Power grid cascading failure risk prediction method and device and storage medium

Abstract

The invention provides a power grid cascading failure risk prediction method and device and a storage medium, and relates to the technical field of power system safety. The power grid cascading failure risk prediction method comprises the steps of obtaining voltage or power flow of a target line; determining risk measurement results of the target line at different time stages by adopting a risk measurement index based on fluctuation of voltage or power flow; and based on the risk measurement result, carrying out weighted accumulation on the voltage or power flow of each time stage, and determining a risk prediction result of the target line. When the risk is predicted, the time period is divided for the voltage or the power flow, and the weighted accumulation is performed on the voltage or the power flow of each time stage based on the risk measurement result of the voltage or the power flow in the time period, so that the contribution degree of the voltage or the power flow data to the line risk is considered from the time dimension and the risk dimension; compared with a risk prediction method which does not divide time periods and equally treats all data in the prior art, the risk prediction method provided by the invention is higher in prediction precision.

Description

Technical Field

[0001] The present invention relates to the field of power system safety technology, and in particular to a power grid cascading failure risk prediction method, device and storage medium. Background Art

[0002] At present, the proportion of renewable energy in urban power grid systems is increasing, and urban power grid systems are facing many challenges. For example, when a traditional generator set in the power grid system fails and the failure is not effectively suppressed, it is very easy to cause a chain failure due to the limited tolerance of renewable energy power generation equipment to low frequencies and weak fault resistance, leading to large-scale power outages. For another example, under some extreme weather conditions, there is a risk that renewable energy power generation equipment will be out of the power grid, which can easily cause a chain failure of the power grid.

[0003] In order to improve the safe and stable operation capability of urban power grids, it is necessary to identify the risks of cascading failures in urban power grids. However, the traditional risk identification of cascading failures in urban power grids has low sensitivity and cannot accurately identify these risks. Therefore, a new type of high-accuracy risk identification method for cascading failures in urban power grids is urgently needed. Summary of the invention

[0004] The present invention provides a method, device and storage medium for predicting the risk of cascading failures in a power grid, so as to solve the defect of insensitivity to the risk identification of cascading failures in the prior art and realize high-precision prediction of the risk of cascading failures in the power grid.

[0005] The present invention provides a method for predicting the risk of power grid cascading failures, comprising the following steps.

[0006] Acquiring parameter information of a target line, the parameter information including: voltage or power flow; the target line includes a potential cascading fault line; Get the voltage or power flow of the target line; Based on the fluctuation of the voltage or the power flow, using a risk measurement index, determining the risk measurement results of the target line at different time stages; Based on the risk measurement result, the voltage or the power flow in each time stage is weighted and accumulated to obtain the risk prediction result of the target line.

[0007] According to a method for predicting the risk of cascading failures in a power grid provided by the present invention, based on the risk measurement result, weighted accumulation is performed on the voltage or the power flow at each time stage to determine the risk prediction result of the target line, including: Taking the risk measurement result as the weight, the voltage or the power flow at each time stage is weighted and nested layer by layer from the time advancement dimension to obtain the risk prediction result of the target line.

[0008] According to a method for predicting the risk of cascading failures in a power grid provided by the present invention, the risk measurement results of each time stage are weighted and nested layer by layer from the time advancement dimension, including: The risk prediction model is used to calculate the risk prediction result of the target route, and the risk prediction model is as follows: in, Indicates the risk prediction results, in chronological order: Indicates the starting time period, and the subscript 1 indicates the ending time period; , , , Respectively Time period, The voltage or current of the time period, the end time period, and the previous time period before the end time period. () represents the risk measurement indicator measurement function.

[0009] According to a method for predicting the risk of cascading failures in a power grid provided by the present invention, the risk measurement index is a CVaR index.

[0010] According to a method for predicting the risk of cascading failures in a power grid provided by the present invention, before acquiring parameter information of a target line, the method further includes: Determining a power grid line flow based on power grid state parameters, wherein the power grid state parameters include load power and energy generation power; Determining a line outage probability based on the line flow; The target line is determined based on the line outage probability.

[0011] According to a method for predicting the risk of cascading failures in a power grid provided by the present invention, the method of determining the power grid line flow based on the power grid state parameters comprises: The grid state parameters are used as input variables, and a random power flow algorithm is used to determine the grid line power flow; the random power flow algorithm is used to determine the mapping relationship between the grid state parameters and the grid line power flow probability distribution.

[0012] According to a method for predicting the risk of cascading failures in a power grid provided by the present invention, before determining the power grid line flow based on the power grid state parameters, the method further includes: Establishing a new energy power generation model; the new energy power generation model describes the power generation of new energy based on simulating the uncertainty of new energy power generation; The renewable energy power generation power is determined based on the renewable energy power generation model.

[0013] According to a method for predicting the risk of cascading failures in a power grid provided by the present invention, the new energy includes photovoltaic new energy and / or wind power new energy; for photovoltaic new energy, the new energy power generation model describes the power generation power based on light intensity; for wind power new energy, the new energy power generation model describes the power generation power based on wind speed.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements any of the above-mentioned methods for predicting the risk of power grid cascading failures.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for predicting the risk of cascading failures of a power grid as described in any one of the above is implemented.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for predicting the risk of cascading failures in a power grid.

[0017] The power grid cascading failure risk prediction method, device and storage medium provided by the present invention divides the voltage or power flow data into time periods, uses risk measurement indicators to measure the risk of voltage or power flow data in different time periods, and performs weighted accumulation of the voltage or power flow data in each time period based on the measurement results, thereby obtaining the final line risk prediction result. Since the present invention divides the basic data for predicting line risks into time stages, and performs weighted accumulation based on the risk measurement results of the data in the time period, that is, the contribution of the data to the final risk prediction result is considered from the time dimension and the risk dimension, compared with the risk prediction method in the prior art that does not divide time periods and treats all data equally, the risk prediction accuracy of the present invention is higher and the sensitivity to risk identification is also higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 This is one of the flow charts of the method for predicting the risk of power grid cascading failures provided in an embodiment of the present invention.

[0020] Figure 2 It is a risk change trend diagram of CVaR based on data at a single time point in an embodiment of the present invention.

[0021] Figure 3 It is a risk change trend diagram of CVaR based on multi-time period data in an embodiment of the present invention.

[0022] Figure 4 This is the second flow chart of the method for predicting the risk of power grid cascading failures provided in an embodiment of the present invention.

[0023] Figure 5 is a probability distribution curve diagram of wind speed provided by an embodiment of the present invention.

[0024] Figure 6 This is a corresponding relationship diagram between the output power of a wind turbine generator and the wind speed provided by an embodiment of the present invention.

[0025] Figure 7 It is a probability distribution curve diagram of light intensity provided by an embodiment of the present invention.

[0026] Figure 8 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Combine the following Figure 1-Figure 7 A method for predicting the risk of cascading failures in a power grid provided by an embodiment of the present invention is described.

[0029] Figure 1 FIG. 1 is one of the flow charts of the method for predicting the risk of power grid cascading failures provided by an embodiment of the present invention, such as Figure 1 As shown, the method includes the following: Step 101: Obtain the voltage or power flow of the target line.

[0030] Step 102: Based on the fluctuation of the voltage or the power flow, a risk measurement index is used to determine the risk measurement results of the target line at different time stages.

[0031] Step 103: Based on the risk measurement result, weighted accumulation is performed on the voltage or the power flow in each time stage to obtain a risk prediction result of the target line.

[0032] In some embodiments, the present invention measures the risk of fluctuations in historical parameter information of the target line by time stages, and performs weighted accumulation of parameter information data of each time period based on the measurement results, that is, the contribution of data to the risk prediction results is considered from the time dimension and the risk dimension. Compared with the traditional risk measurement prediction method that does not divide the time period, the risk prediction accuracy of the present invention is higher and the sensitivity to risk identification is also higher.

[0033] It should be noted that the target line in this embodiment can come from a target line set, which can be a set of all lines in the power grid, and can also be understood as a set of potential cascading failure lines in the power grid, and the potential cascading failure lines can be understood as pre-screened lines with a certain cascading failure risk (which will be further introduced later).

[0034] When the target line set is the set of all lines in the power grid, we can traverse all lines in the target line set when performing risk prediction, that is, execute steps 101 to 103 for all lines respectively, that is, the risks of all lines in the power grid are accurately calculated and predicted, which can reflect the risk status of the power grid more completely and without omission.

[0035] In some embodiments, step 103 performs weighted accumulation of the voltage or the power flow at each time stage based on the risk measurement result to determine the risk prediction result of the target line. Further steps may be as follows: Taking the risk measurement result as the weight, the voltage or the power flow at each time stage is weighted and nested layer by layer from the time advancement dimension to obtain the risk prediction result of the target line.

[0036] Specifically, the above-mentioned layer-by-layer weighted nested calculation can be understood as using the following risk prediction model for calculation, which is of course not limited to the following model form.

[0037] in, Indicates the risk prediction results, in chronological order: Indicates the starting time period, and the subscript 1 indicates the ending time period; , , , Respectively Time period, The voltage or current of the time period, the end time period, and the previous time period before the end time period. () represents the risk measurement index measurement function , βIt is a nonlinear weight parameter, which aims to amplify the impact of large deviations, and is used to perform nonlinear weighting on important factors in the risk of power grid operation, so as to more accurately reflect the characteristics of nonlinear losses and risks when the deviation in the power grid increases. The risk measurement indicator can be a CVaR indicator or a VaR indicator. , express t The confidence level for the time period (such as 95% or 99%) is used to determine the calculation parameters of the risk measurement function (CVaR or VaR).

[0038] That is to say, if the risk of the current time period of the line is predicted, it is necessary to obtain the historical data of parameter information of several time periods before the current time period. The earliest time period of these time periods is the T time period mentioned above, and the current time period is the terminal time period mentioned above. In terms of the order of time advancement, the risk prediction results of the latter time period are related to the risk measurement results of the previous time period.

[0039] It should be noted that the time period in the embodiment of the present invention is a ms-level time period. For example, 1ms is a time period. When predicting the risk result at t=10ms, voltage or power flow fluctuation data of several historical time periods such as t=5ms, t=6ms, t=7ms, t=8ms, and t=9ms are required. After substituting the above data into the risk prediction model, the risk prediction model becomes .

[0040] The embodiment of the present invention considers the contribution of data to the final risk prediction results from the time dimension and the risk dimension, which is in line with actual laws. Based on this, compared with the traditional risk prediction method that regards the influence of data in all time periods on the risk prediction results as equally important, the present invention can better reflect objective laws and thus predict risks more accurately.

[0041] This embodiment divides the entire time series data into multiple sub-intervals or stages so that the risk changes in each time period can be analyzed in more detail. In each sub-interval, the risk measurement index value is calculated to evaluate potential extreme loss situations, focusing on tail risks. The initial return value of each stage is extracted and placed in an array as a benchmark for recursive nesting. Using a recursive formula, the risk measurement index values ​​of each stage are nested and superimposed layer by layer to generate a cumulative risk measurement for the entire time period, that is, the above-mentioned risk prediction result. The trend of the cumulative multi-stage risk measurement index values ​​changing over time is plotted into a graph to help observe dynamic changes, identify extreme fluctuations and tail risks, and to review and fine-tune the results after calculation.

[0042] In some embodiments, the risk measurement index in step 102 can be selected from the expectation-variance risk measurement functional, the expectation-bias risk measurement functional, the VaR risk measurement functional, and the CVaR risk measurement functional. In an embodiment of the present invention, the above-mentioned risk measurement index is applied to the distribution network, the risk brought by uncertainty is quantified, and it is used as an indicator for optimizing the distribution network to measure the loss risk of the uncertainty of new energy generation on the operating performance of the distribution network, that is, the risk cost loss during the operation process, thereby improving the operating efficiency of the distribution network and enhancing the stability and economy of the system.

[0043] Expectation-variance risk measure functional: in, represents the expected value of the random variable Z, ≥0 is a given coefficient used to adjust the relative importance of expectation and variance in risk measurement. = 0, the risk measure depends only on the expected value; when When >0, the variance plays a role in risk measurement, and the larger it is, the greater the impact of the variance on risk measurement. Represents the variance of the random variable Z. Expected-upper half deviation risk measure functional: in, It means that under probability P, The Lp norm of the upper part (i.e. the part where Z is greater than the expected rate of return E(Z)). The Lp norm is a method of measuring the size of a vector, where p is a given positive integer. The specific definition is: .

[0044] VaR risk measurement functional: Among them, Fz is the cumulative probability density function of the random loss variable Z. α is the confidence level, which means that under this probability, the loss of the investment portfolio will not exceed the VaR value.

[0045] CVaR risk measure functional: in, is the confidence level, The coefficient plays a role of normalization and weight adjustment in the CVaR formula, ensuring that the calculated conditional risk value is a reasonable average loss value and matches the given confidence level.

[0046] By nesting risk functionals to assess risks in multi-stage decision-making problems, the risk changes of the system in multiple time stages are taken into account, and the characteristics of time correlation and stage changes can be captured. This method nests the risk measures of each period layer by layer through a function to obtain a comprehensive risk assessment of the entire decision-making process. The nested structure means that the risk assessment of the subsequent stage will be based on the assessment results of the previous stage. The core is to dynamically update the assessment and decision-making strategy of future risks based on existing information at each time point.

[0047] In some embodiments of the present invention, the risk measurement indicator is selected from the CVaR risk functional (conditional value at risk). The prediction results of the risk prediction model based on the CVaR risk functional are hereinafter referred to as dynamic multi-stage CVaR results.

[0048] The dynamic multi-stage CVaR results are more suitable as a risk assessment tool, and the dynamic multi-stage CVaR can more accurately capture the fluctuations at different stages. When the risk fluctuations at each stage are significant, the dynamic multi-stage CVaR captures the risk accumulation effect through nested calculations, thereby more sensitively reflecting the dynamic changes of risks, especially in the case of large fluctuations. Figure 2 and Figure 3 It is a risk change trend chart based on a set of data, using CVaR at a single time point and multiple time periods. Unlike CVaR at a single time point, dynamic multi-stage CVaR can track the risk transmission effect between multiple time periods. It takes into account the relationship between each stage, so that the risk assessment is no longer an isolated single-time point measurement, but combines historical information and predicts the future, which is more in line with the actual situation. For the risk measurement of extreme events, dynamic multi-stage CVaR has more advantages because it is based on conditional value at risk (CVaR), is more sensitive to tail risk (extreme losses), and can therefore better reflect potential risks. Figure 2 Figure 3 It can be seen from the display that the dynamic multi-stage CVaR can rise in time and provide a higher risk warning when extreme risks occur. Compared with the single-stage risk measurement, the dynamic multi-stage CVaR shows a smoother trend, which helps to identify long-term risk trends rather than relying solely on short-term fluctuations. In addition, the dynamic multi-stage CVaR can also provide a more stable risk view, which is convenient for formulating long-term risk management strategies. It can be seen that the dynamic multi-stage CVaR proposed in the embodiment of the present invention can more effectively integrate information from multiple time stages, capture the gradual accumulation and transfer effects of risks, and is a more comprehensive and flexible risk measurement tool.

[0049] It is understandable that the single time point mentioned above refers to a certain time point, and the multiple time periods refer to multiple time periods obtained by subdividing the data of a certain time point in time. For example, the single time point is the time point t=1ms, and its corresponding data is the data from t=0ms to t=1ms. At this time, the multiple time periods refer to the subdivision of t=0ms to t=1ms in time, such as subdividing into 5 time periods of t=0ms to t=0.2ms, t=0ms to t=0.4ms, t=0ms to t=0.6ms, t=0ms to t=0.8ms, and t=0ms to t=1ms.

[0050] The above-mentioned dynamic multi-stage risk prediction method provided by the embodiment of the present invention takes into account the changing characteristics of risks at different time stages. By introducing the time dimension into risk analysis, it can more accurately reflect the risk status of the system at different times, and improve the timeliness and accuracy of risk assessment. At the same time, it can capture the gradual diffusion process of chain failures in the system. By dividing the fault events into multiple stages for analysis, the fault propagation path and its impact on the system as a whole can be more clearly revealed, which facilitates the identification of high-risk nodes and key fault points. In addition, it can take into account both local and global risk distribution, and can perform comprehensive analysis at different levels, thereby providing a more targeted risk control strategy.

[0051] It is mentioned above that the target line can come from a target line set, and the target line set can be a set of potential cascading failure lines, and the potential cascading failure lines can be understood as lines that have been pre-screened and have a certain cascading failure risk. The screening of potential cascading failure lines is introduced below. In this embodiment, the following steps may be specifically included: Step 201: determining a power grid line flow based on power grid state parameters, wherein the power grid state parameters include load power and energy generation power; Step 202: Determine the line outage probability based on the line flow; Step 203: Based on the line outage probability, determine the target line. The target line here is equivalent to the potential cascading failure line that has been pre-screened.

[0052] In some embodiments of the present invention, step 201 may specifically include: using a grid state parameter as an input variable and using a random power flow algorithm to determine a grid line power flow.

[0053] In the process of analyzing line faults, in order to obtain the probability distribution of line power flow, a random power flow algorithm is usually used. The SRSM method is a commonly used probability analysis method with the characteristics of simple operation and fast speed. The core idea is to construct a polynomial model of the input-output relationship and determine the coefficients in the polynomial with a small amount of sampling to obtain the probability distribution of the output response. When calculating cascading failures, multiple power flow analyses are required, and the basic process of the SRSM method can be summarized as follows. First, SRSM generally transforms the input random variable I into a standard normal distribution variable. The transformation process is: In the formula is the cumulative distribution function, is the cumulative probability distribution function of the normal distribution, is an n-dimensional standard normal distribution variable.

[0054] Establish a polynomial. For the output variable , the remaining input variables The relationship can be expressed as: In this method, Hm(ξ) is the m-th order Hermite polynomial, The coefficients are. Increasing the order of the polynomial can effectively improve the accuracy of the simulation, but this will also increase the number of coefficients to be estimated accordingly, thereby requiring more sampling times to determine these coefficients. According to research, the second-order polynomial expansion is sufficient to achieve satisfactory accuracy. In addition, the cross terms in the second-order polynomial have a relatively limited contribution to improving accuracy, so these cross terms can usually be ignored. Therefore, the present embodiment adopts a second-order polynomial expansion that ignores the cross terms.

[0055] In this embodiment, the renewable energy power generation and load output together constitute the output variable I. Finally, by sampling the standardized input variables, random variable samples can be generated according to the above formula, and the probability distribution of the output response can be obtained through statistical analysis of these samples.

[0056] The random power flow calculation of the line can be realized by the Newton-Raphson method (Newton-Raphson method for short). This method can effectively solve the phase angle and amplitude of the node voltage. The active power flow equation can be expressed as: Among them, the variable and Respectively represent the two nodes of the line (the nodes here refer to the power supply nodes), and are the conductance and susceptance of the line.

[0057] When considering the uncertainty of renewable energy output and load power, renewable energy power generation and load power are used as input random variables. Using the SRSM method, a mapping relationship between input variables and branch active power flow can be established to quickly obtain the probability distribution of branch power flow. According to the above formula, the second-order polynomial of the line power flow expansion is: When making risk predictions, a series of constraints need to be met to ensure the normal operation of the power system. To this end, the operating status of the distribution network must first be determined through power flow calculations, so as to further obtain key indicators such as voltage deviation.

[0058] 1) Power balance constraint: During normal operation, the distribution network must always maintain a balance between the total system power and the total load. Therefore, power balance constraints must be added to the model to ensure the balance between supply and demand in the system.

[0059] in, , , , They represent the active power rate of node n, the injected wind power generation active power and photovoltaic power generation active power, and the active load at the node respectively. , , , They represent the reactive power of the node, the injected wind power reactive power, the photovoltaic power reactive power, and the reactive load at the node respectively. is the voltage at node n, is the voltage at node q. The total number of nodes in the distribution network system is n. , , are the line conductance, susceptance and voltage phase angle between system nodes n and q respectively.

[0060] 2) Node voltage limit: The normal operation of the distribution network system requires that the node voltage remain within a specific upper and lower limit. When the voltage of a node in the system exceeds this limit, the stability of the system will be affected, increasing the risk of failure. Therefore, in order to ensure system safety, the system operation must be controlled within the specified voltage upper and lower limits.

[0061] 3) Power output constraints: There are upper and lower limits on the output of wind power generation and photovoltaic power generation.

[0062] 4) Branch power constraint: There is a maximum limit on the power flowing through a branch.

[0063] 5) Network constraints: The distribution network needs to operate according to a radial topology. There must be no loops and no isolated nodes that are not connected to the distribution network.

[0064] In some embodiments of the present invention, step 202 and step 203 are specifically implemented by the following method: A cascading failure event can be represented as the set , where each A represents a group of component failures, that is , from the state Transfer to state The probability is: In the formula, is the failure probability of line c in the jth stage, Represents the normal line collection at stage j.

[0065] Under extreme weather conditions, the withdrawal of renewable energy is closely related to the incidence of cascading failures, especially in terms of voltage. When a cascading failure occurs due to the withdrawal of renewable energy, the heat generation of line c increases significantly, which can easily lead to a short circuit. In addition, when the load is too high, the voltage will approach its stability limit, further causing line failures. This paper models the failure rate of the component as a piecewise function of the load rate, and the relationship between the two is: in, is the load factor of line c, is the load rating, is the load rate limit. According to the above formula, the line failure rate under deterministic operation can be expressed as a piecewise function of the load rate, and the relationship between the two is: Considering the power probability density in the line after the fault occurs and the line failure probability under the deterministic operation mode, the outage probability of line c under this condition is: in, represents the number of discrete intervals of the probability density function, represents the average load rate in each interval. Considering the probability and consequences of cascading failures, a cascading failure risk index is proposed. The probability of cascading failure can be expressed as: In this embodiment, after obtaining the probability of cascading failures of the accident chain, the high-risk cascading failure lines are further screened based on the magnitude of the failure probability, and the set of screened high-risk cascading failure lines is the target line set mentioned above. This is equivalent to first determining the outage probability based on the line flow, and then preliminarily determining the high-risk cascading failure lines, that is, the potential cascading failure lines mentioned above (preliminary identification), and then measuring the parameter information of the potential cascading failure lines using risk measurement indicators, and determining the final high-risk cascading failure lines based on the measurement results (precise identification). The above-mentioned preliminary identification combined with precise identification has the following advantages: The SRSM method in the preliminary identification stage is an efficient calculation method. It quickly evaluates line power flow and outage probability through probabilistic approximation, and can quickly and effectively filter out most low-risk lines and identify high-risk lines, thereby focusing the precise analysis in the precise identification stage on the high-risk candidate lines (i.e., potential cascading failure lines) that really need attention, reducing the amount of calculation in the second stage and greatly improving the analysis efficiency. This not only saves resources, but also makes the entire analysis process more targeted.

[0066] Managers can also set priorities based on the preliminary screening results and implement risk control measures more effectively.

[0067] In the aforementioned embodiment of the present invention, a scheme that does not include the preliminary identification stage is also mentioned. In this scheme, the potential cascading fault line can be understood as any line in the power grid, and the target line set can be understood as the set of all lines in the power grid. The advantage of this scheme is that each line is deeply analyzed without omissions, and the risk status of the power grid can be more completely reflected. However, due to its large resource consumption, it is mostly used in situations with sufficient resources or small-scale store systems.

[0068] After the high-risk lines and their risk levels are predicted in the above embodiments, targeted risk control can be performed on the high-risk lines. As for the specific control method, the existing control strategy can be adopted, and the present invention will not elaborate on it or limit it.

[0069] Regarding the renewable energy power generation in step 201 above, some embodiments of the present invention provide a calculation model for the renewable energy power generation in combination with the uncertainty of renewable energy power generation.

[0070] The calculation model of wind power renewable energy power generation is introduced below.

[0071] Wind power generation is integrated into the distribution network, usually in the form of a wind farm to supply power to the distribution network and input power to the grid. The output of a wind turbine can be expressed as: Where ρ is the density of air, is the unit area perpendicular to the wind speed direction, is the wind energy utilization coefficient.

[0072] The output power of wind turbines is mainly determined by wind speed, which is very random and difficult to accurately model. A probability model is needed to describe the uncertainty of wind speed. Weibull distribution is used to describe wind speed, and the probability density function is: k and c are the shape parameter and scale parameter of Weibull distribution respectively. The mean and standard deviation of wind speed are obtained by statistics of local wind speed history data, and k and c are calculated by the mean and standard deviation. Figure 5 It is the probability distribution curve of wind speed, which represents the probability distribution of wind speed in a certain period of time: The cumulative probability distribution function (CPDF) of wind speed is used for probability sampling and is given by the following formula: In order to facilitate calculation, the relationship between the output power of wind turbines and wind speed is linearized and approximated, and the output power of wind power generation is divided into four stages and three working states according to the wind speed. The relationship can be expressed as follows: in, Indicates the rated output power of the wind turbine. Indicates rated wind speed.

[0073] Figure 6 The corresponding relationship between the output power of the wind turbine and the wind speed is shown in Figure 2. According to the probability density function of wind speed and the relationship between the output power of the wind turbine and wind speed, the probability distribution of the output active power can be calculated using probability statistics knowledge. When the wind speed is between the cut-in and the rated value, the probability equation for wind power generation is: The corresponding reactive power output of wind power generation is: The calculation model of wind power renewable energy power generation is introduced below.

[0074] The output power of photovoltaic power generation is approximated as follows: in, represents the conversion efficiency of photovoltaic cells, Represents the photosensitive area of ​​the photovoltaic panel, Indicates the light intensity per unit area at a certain moment.

[0075] Light intensity is similar to wind speed, and is also affected by natural weather factors, with strong random uncertainty. The random probability model of light intensity is described as Beta distribution, and the probability density function is as follows. The corresponding light intensity probability distribution curve is as follows: Figure 7 .

[0076] Among them, α and β are the shape parameters of the Beta distribution function. The mean and standard deviation of the local light intensity can be obtained through the historical data of light intensity in the photovoltaic power generation area. The calculation formula is as follows. The mean of the local light intensity can be obtained through the historical data of light intensity in the photovoltaic power generation area. and standard deviation .

[0077] The cumulative probability distribution function (CPDF) of light intensity is used for probability sampling, and the formula is as follows: There is a mathematical relationship between photovoltaic power output and light intensity as follows: Where E is the threshold of light intensity.

[0078] According to the above formula, the probability density function of photovoltaic power output power can be derived as follows: Similar to wind power generation, photovoltaic power generation also uses constant power factor control and is also considered as a PQ node when it is connected to the distribution network. The reactive power output of photovoltaic power generation can be calculated using the following formula.

[0079] The LHS method is an efficient random sampling technique first proposed by McKay in 1979. Its characteristics are that it can perform stratified random sampling with equal probability within the distribution interval, and has the advantages of uniform stratification and full sample coverage. Even with a small number of sampling times, it can ensure that the tail and marginal samples are not missed, which enhances the comprehensiveness and representativeness of the sampling.

[0080] The above calculations of the probability density function (PDF) and cumulative distribution function (CDF) of wind speed and light intensity provide a distribution basis for sampling. The specific Latin hypercube sampling steps are as follows: Set range and plot: Set the range of wind speed and light intensity, and reverse the points in the cumulative distribution to obtain the corresponding sample values.

[0081] Sampling preprocessing: Set the total number of samples N, determine the number of intervals to split the CDF, calculate the midpoint of each interval, and ensure that the samples are evenly distributed throughout the range.

[0082] Execute sampling: According to the distribution functions of wind speed and light intensity, inversely calculate the light intensity and wind speed sample values ​​corresponding to the midpoint of CDF.

[0083] Power output model: The power output of the wind turbine is calculated based on the sampled wind speed, the power variation law under different wind speeds is considered, and the power output of the photovoltaic cell is calculated using the light intensity.

[0084] Figure 8 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 8 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820 and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the power grid cascading failure risk prediction method, the method comprising: obtaining parameter information of the target line, the parameter information comprising: voltage or flow; the target line comprising a potential cascading failure line; based on the fluctuation of the parameter information, using a risk measurement index to determine the risk measurement results of the target line at different time stages; weighted accumulation of the risk measurement results of each time stage to determine the risk prediction result of the target line; adjusting the risk control strategy based on the risk prediction result.

[0085] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0086] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the power grid cascading failure risk prediction method provided by the above-mentioned methods, and the method includes: obtaining parameter information of the target line, and the parameter information includes: voltage or current; the target line includes a potential cascading failure line; based on the fluctuation of the parameter information, using a risk measurement indicator to determine the risk measurement results of the target line in different time stages; weightedly accumulating the risk measurement results of each time stage to determine the risk prediction results of the target line; adjusting the risk control strategy based on the risk prediction results.

[0087] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the power grid cascading failure risk prediction method provided by the above-mentioned methods, the method comprising: obtaining parameter information of a target line, the parameter information comprising: voltage or current; the target line comprising a potential cascading failure line; based on the fluctuation of the parameter information, using a risk measurement index to determine the risk measurement results of the target line at different time stages; weightedly accumulating the risk measurement results of each time stage to determine the risk prediction result of the target line; and adjusting the risk control strategy based on the risk prediction result.

[0088] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0089] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the risk of cascading failures in a power grid, characterized in that: include: Obtain the voltage or power flow of the target line; Based on the fluctuation of the voltage or the power flow, using a risk measurement index, determining the risk measurement results of the target line at different time stages; Based on the risk measurement result, the voltage or the power flow in each time stage is weighted and accumulated to obtain the risk prediction result of the target line.

2. The method for predicting the risk of power grid cascading failures according to claim 1, characterized in that: Based on the risk measurement result, weighted accumulation is performed on the voltage or the power flow in each time stage to determine the risk prediction result of the target line, including: Taking the risk measurement result as the weight, the voltage or the power flow at each time stage is weighted and nested layer by layer from the time advancement dimension to obtain the risk prediction result of the target line.

3. The method for predicting the risk of cascading failures in a power grid according to claim 2, characterized in that: The step of performing layer-by-layer weighted nested calculation of the voltage or the power flow at each time stage from the time advancement dimension includes: The risk prediction model is used to calculate the risk prediction result of the target route, and the risk prediction model is as follows: in, Indicates the risk prediction results, in chronological order: Indicates the starting time period, and the subscript 1 indicates the ending time period; , , , Respectively Time period, The voltage or current of the time period, the end time period, and the previous time period before the end time period. () represents the risk measurement indicator measurement function.

4. The method for predicting the risk of cascading failures in a power grid according to any one of claims 1 to 3, characterized in that: The risk measurement indicator is the CVaR indicator.

5. The method for predicting the risk of power grid cascading failures according to claim 1, characterized in that: Before acquiring the parameter information of the target line, the method further includes: Determining a power grid line flow based on power grid state parameters, wherein the power grid state parameters include load power and energy generation power; Determining a line outage probability based on the line flow; The target line is determined based on the line outage probability.

6. The method for predicting the risk of cascading failures in a power grid according to claim 5, characterized in that: The determining of the power grid line flow based on the power grid state parameter comprises: The grid state parameters are used as input variables, and a random power flow algorithm is used to determine the power line flow of the grid; the random power flow algorithm is used to determine the mapping relationship between the grid state parameters and the probability distribution of the power line flow of the grid.

7. The method for predicting the risk of power grid cascading failures according to claim 5 or 6, characterized in that: Before determining the power grid line flow based on the power grid state parameters, the method further includes: Establishing a new energy power generation model; the new energy power generation model describes the power generation of new energy based on simulating the uncertainty of new energy power generation; The renewable energy power generation power is determined based on the renewable energy power generation model.

8. The method for predicting the risk of cascading failures in a power grid according to claim 7, characterized in that: The new energy includes photovoltaic new energy and / or wind power new energy; for photovoltaic new energy, the new energy power generation model describes the power generation based on light intensity; for wind power new energy, the new energy power generation model describes the power generation based on wind speed.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the program, the method for predicting the risk of cascading failures in a power grid as described in any one of claims 1 to 8 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the risk of cascading failures in a power grid as claimed in any one of claims 1 to 8 is implemented.

Images