Distribution network key element reliability parameter correction method for distributed power supply access

Abstract

The invention discloses a distributed power supply access distribution network key element reliability parameter correction method, and relates to the technical field of distribution network key element reliability parameter correction, and the method comprises the steps: collecting power distribution network data, and carrying out the modeling of fan output and photovoltaic output; modeling is carried out on the output uncertainty influence of the distributed power supply by utilizing a Songnoi polyhedron, and a Monte Carlo method is improved to carry out key element identification; and selecting a BPNN as a neural network model, taking power distribution network data as a training set and carrying out normalization processing, training the BPNN model and then carrying out reliability parameter correction. According to the method, the management of the volatility and uncertainty of the distributed power supply is enhanced through a Chicken polyhedron method, the overall operation efficiency and reliability of the power distribution network are improved, the reliability parameters of key elements of the power distribution network are allowed to be dynamically corrected by adopting the BPNN model, the operation flexibility and toughness of the power distribution network are improved, and the reliability of the power distribution network is improved. And meanwhile, stable operation of a power grid and power supply reliability are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of reliability parameter correction of key components of a distribution network, and in particular to a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected. Background Art

[0002] In today's power system, the access of distributed power sources (such as wind power and solar energy) is increasing, which poses new challenges to the stable operation and reliability of the power grid. The volatility and uncertainty of distributed power sources increase the complexity of distribution network operation. Therefore, accurately evaluating and correcting the reliability parameters of key components of the distribution network has become an important means to ensure the stable operation of the power grid. In recent years, distribution network reliability assessment and parameter correction methods for distributed power access have been widely studied. These methods collect distribution network operation data and use various models to predict the output of distributed power sources such as wind turbines and photovoltaics, thereby evaluating the reliability of the power grid and correcting the reliability parameters of key components to meet the actual needs of power grid operation.

[0003] Although the existing technology can deal with the impact of the uncertainty of distributed power generation on the reliability of the power grid to some extent, there are still some shortcomings. First, many methods rely on overly simplified models when dealing with the uncertainty of distributed power generation output, which makes it difficult to accurately capture the complexity of wind and solar power output. Second, the existing technology often ignores the impact of environmental factors and operating conditions on the reliability parameters of key components of the distribution network, resulting in insufficient accuracy of reliability parameter correction. In addition, the existing methods often lack effective data processing and learning mechanisms in the parameter correction process, and cannot make full use of the historical operation data of the distribution network for in-depth analysis and learning, thus affecting the accuracy and adaptability of the correction results. By using Zeno polyhedron to model the impact of the uncertainty of distributed power generation output and improving the Monte Carlo method for key component identification, this method can more accurately identify the key factors affecting the reliability of the power grid. Finally, BPNN is selected as the neural network model to conduct in-depth learning and analysis of the distribution network data to realize the dynamic correction of reliability parameters. This method not only considers the uncertainty of distributed power generation, but also combines the actual operating conditions and historical data of the distribution network. Through advanced data processing and learning mechanisms, the accuracy and efficiency of reliability parameter correction are improved, providing strong technical support for the stable operation of the distribution network. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing method for correcting the reliability parameters of key distribution network components has low reliability, low accuracy and low efficiency, and the problem of how to improve the adaptability of the distribution network to the access of renewable energy and the reliability of the overall operation.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for correcting reliability parameters of key components of a distribution network with distributed power sources connected, comprising collecting distribution network data and modeling wind turbine output and photovoltaic output; using Zeno polyhedron to model the influence of uncertainty on distributed power output and improving the Monte Carlo method for key component identification; selecting BPNN as a neural network model, using distribution network data as a training set and normalizing it, and correcting reliability parameters after training the BPNN model.

[0007] As a preferred solution of the method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected as described in the present invention, the distribution network data includes distribution network components, distribution network structure, distribution network environmental operating parameters, historical fault data, and reliability evaluation indicators.

[0008] As a preferred solution of the reliability parameter correction method of key components of the distribution network connected to the distributed power supply of the present invention, wherein: the modeling of wind turbine output and photovoltaic output includes adopting an autoregressive moving average model to model and predict wind speed, collecting regional wind speed data, and based on the original wind speed data v t Calculate the average wind speed μ and the standard deviation σ of the wind speed distribution, and t After standardization, the original data y t With v t The relationship is expressed as:

[0009]

[0010] The autoregressive moving average model is expressed as:

[0011] y t =a1y t-1 +a2y t-2 +…+a m y t-m +α t +c1α t-1 +c2α t-2 +…c n α t-n

[0012] Among them, t represents the time point, m represents the order of the autoregressive part, and a m represents the autoregressive coefficient, c n represents the moving average coefficient, α t is a normally distributed white noise, and the functional relationship between the output power P1 of the wind turbine and the wind speed V is expressed as:

[0013]

[0014] Among them, V a Indicates the cut-in wind speed of the wind turbine, V b Indicates the rated wind speed of the wind turbine, V c Indicates the cut-out wind speed of the wind turbine, P e Represents the rated power of the wind turbine generator, and a, b, and c are calculated as follows:

[0015]

[0016] Among them, a is the initial slope influence coefficient of the wind turbine output power curve, b is the slope influence coefficient of the power curve when the wind speed increases moderately, and c is the slope influence coefficient of the power curve at the rated wind speed.

[0017] As a preferred solution of the method for correcting reliability parameters of key components of the distribution network connected to the distributed power source of the present invention, wherein: the modeling of the wind turbine output and the photovoltaic output also includes using software to establish a light intensity model to obtain a solar radiation intensity model of the region, the output power of the photovoltaic power generation system is positively correlated with the solar radiation intensity, and the output model of the photovoltaic power generation system is expressed as:

[0018] P2=rAρK

[0019] Among them, P2 is the photovoltaic power output power, r represents the solar radiation intensity, A represents the photovoltaic array area, ρ represents the conversion efficiency of photovoltaic modules to convert light energy into electrical energy, and K represents the correction factor.

[0020] As a preferred solution of the method for correcting the reliability parameters of key components of the distribution network connected to the distributed power supply of the present invention, the modeling of the influence of the uncertainty of the output of the distributed power supply includes that each node in the distribution network has a maximum acceptable capacity. When a fault occurs, the circuit breaker is activated, the topology of the distribution network changes, and the node voltage is raised by the distributed power supply, resulting in an increase in the node voltage. The real-time voltage deviation does not exceed ±7% of the nominal voltage. The uncertainty of the output of the distributed power supply is modeled using Zeno polyhedron, which is expressed as:

[0021]

[0022] in, represents the center of the Zeno polyhedron, located in n-dimensional real space, represents the generating vector of the Zeno polyhedron, p is the number of generating vectors, β i is the scaling factor for the i-th generated vector.

[0023] As a preferred solution of the reliability parameter correction method of key components of the distribution network connected to the distributed power supply of the present invention, the key component identification includes that the trouble-free operation time and fault repair time of the system components are subject to exponential distribution, and the sampling form of the trouble-free operation time and fault repair time of a single component is expressed as:

[0024]

[0025] Among them, TTF represents the trouble-free operation time of the component, TTR represents the fault repair time of the component, λ represents the failure rate of the component, δ represents the repair rate of the component, MTTF represents the mean failure time of the component, MTTR represents the mean repair time of the component, R1 and R2 represent random numbers with a value interval of (0,1), and the key components of the distribution network are identified by the improved Monte Carlo simulation method.

[0026] Read the original parameters of the distribution network and initialize the data, determine that the initial key component set includes all distributed power sources, generate a set of random numbers with the same number of elements as the initial key component set, and the value range of the random numbers is between (0,1). The fault-free working time and fault repair time of the initial key components are calculated based on the random numbers in the sampling form of the fault-free operation time and fault repair time of a single component. The distribution network only includes one-dimensional faults of the components in the initial key component set, and the components outside the set operate normally. The system state and the corresponding duration are obtained, the power flow after the system fault is calculated, the node voltage of each node after the fault is determined, and it is judged whether the distributed power source is offline. If the distributed power source is offline, the overvoltage impact range caused by the fault is determined. If the distributed power source is not offline, the operating state of each load point is determined.

[0027] When the system running time is less than the simulation time, a set of random numbers with the same number of elements as the initial key element set is regenerated.

[0028] When the system operation time is greater than or equal to the simulation time, the number of component failures and failure time during the simulation period are counted, the system reliability index is calculated, and the key component set is obtained based on component failures.

[0029] As a preferred solution of the reliability parameter correction method of key components of the distribution network connected to the distributed power source of the present invention, the reliability parameter correction includes selecting BPNN as the neural network model, using the distribution network data as the training set and normalizing it, and correcting the reliability parameters after training the BPNN model; the input vector X of the BPNN n It includes the initial reliability parameters of key components, operating environment parameters and reliability indicators calculated based on system historical data, and the output vector Y m is the reliability parameter of the key component after correction; before training BPNN, X nThe parameters in have different dimensions, including the average power outage frequency index, the average power outage duration index, the expected value of insufficient power, and the ambient temperature and humidity in the system reliability index. Based on the convergence of BPNN training, the dimension differences between different types of variables are eliminated, and the input sample data is normalized and preprocessed, which is expressed as:

[0030]

[0031] Among them, X 0 Represents the normalized BPNN input vector; the forward propagation process of BPNN from input to output is expressed as:

[0032] Y 0 =f L (…f l (…f2(X 0 )

[0033]

[0034] Among them, Y 0 represents the normalized output vector, L represents the total number of layers of BPNN, and Represent the input vector and output vector of the lth layer respectively, f l (.) represents the corrected output function of the lth layer, f L (.) represents the modified output function of the total number of layers of BPNN, s(.) represents the activation function of the neuron, and W l-1 represents the weight between the l-1th layer and the lth layer, b l Represents the bias of the lth layer, selects the rectified linear unit function as the activation function of the hidden layer and the output layer, and calculates the activation function as:

[0035]

[0036] The modification of weight W and bias b during back propagation is expressed as:

[0037]

[0038]

[0039]

[0040]

[0041]

[0042] Among them, E represents the error, N represents the number of samples in the training set, O and O iThey represent the target value of neural network training and the actual value of the i-th iteration respectively, η represents the learning rate, is the loss function E relative to the weight W k The gradient of is the loss function E relative to the bias b k The gradient of the system is used to calculate the reliability index of the historical data as the training target value, W k and b k Respectively represent the weight and bias of the output layer; after the BPNN training results meet the accuracy requirements, the BPNN output vector is denormalized when calculating the revised key component reliability parameters, which is expressed as:

[0043] Y m =Y 0 (max(Y 0 )-min(Y 0 ))+min(Y 0 )

[0044] Output vector Y m The revised reliability parameter results of key components.

[0045] Another object of the present invention is to provide a reliability parameter correction system for key components of a distribution network with distributed power sources connected thereto, which can model the influence of uncertainty on the output of distributed power sources by using Zeno polyhedrons, improve the Monte Carlo method for key component identification, and solve the problem of low reliability in the current reliability parameter correction of key components of the distribution network.

[0046] As a preferred solution of the reliability parameter correction system of key components of the distribution network connected to the distributed power source described in the present invention, it includes: a data processing module, a component identification module, and a parameter correction module; the data processing module is used to collect distribution network data and model the wind turbine output and photovoltaic output; the component identification module is used to use Zeno polyhedron to model the influence of uncertainty on the output of distributed power sources, and improve the Monte Carlo method to identify key components; the parameter correction module is used to select BPNN as the neural network model, use the distribution network data as a training set and normalize it, and perform reliability parameter correction after training the BPNN model.

[0047] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a method for correcting reliability parameters of key components of a distribution network for access of distributed power sources.

[0048] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected.

[0049] Beneficial effects of the present invention: The reliability parameter correction method of key components of the distribution network connected to distributed power sources provided by the present invention predicts the wind speed by adopting an autoregressive sliding average model, and can accurately simulate the dynamic relationship between the output power of the wind turbine and the wind speed. In addition, the light intensity model established by the software can accurately predict the output power of the photovoltaic power generation system. Taking into account the changes in solar radiation intensity, the efficiency and reliability of the access of distributed power sources to the distribution network are improved, and the impact of renewable energy fluctuations on the stability of the distribution network is reduced. Through the Zeno polyhedron method, the impact of the output uncertainty of distributed power sources on the distribution network can be finely simulated and analyzed. The improved Monte Carlo method further improves the accuracy of identifying key components of the distribution network, enhances the management of the volatility and uncertainty of distributed power sources, and improves the overall operation efficiency and reliability of the distribution network. The BPNN model allows dynamic correction of the reliability parameters of key components of the distribution network, reflects the real-time changes in the operation and environmental conditions of the distribution network, realizes dynamic reliability parameter correction based on real-time data, improves the flexibility and resilience of the operation of the distribution network, and ensures the stable operation of the power grid and the reliability of power supply. The present invention achieves better results in terms of efficiency, reliability and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0051] Figure 1 The present invention provides an overall flow chart of a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected, provided in the first embodiment of the present invention.

[0052] Figure 2 A distribution network structure diagram of a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected, provided in accordance with a second embodiment of the present invention.

[0053] Figure 3 A comparison chart of the system average power outage frequency index of a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected, provided in a second embodiment of the present invention.

[0054] Figure 4 A comparison chart of system average power outage duration indicators of a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected, provided in a second embodiment of the present invention.

[0055] Figure 5A comparison diagram of expected values ​​of system power shortage for a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected, provided in a second embodiment of the present invention.

[0056] Figure 6 The third embodiment of the present invention provides an overall flow chart of a system for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected. DETAILED DESCRIPTION

[0057] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0058] Example 1

[0059] Reference Figure 1 , which is an embodiment of the present invention, provides a method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected, comprising:

[0060] S1: Collect distribution network data and model wind turbine output and photovoltaic output.

[0061] Furthermore, the distribution network data includes distribution network components, distribution network structure, distribution network environmental operating parameters, historical fault data, and reliability assessment indicators.

[0062] It should be noted that the modeling of wind turbine output and photovoltaic output includes the use of autoregressive moving average model to model and predict wind speed, collect regional wind speed data, and use the original wind speed data v t Calculate the average wind speed μ and the standard deviation σ of the wind speed distribution, and t After standardization, the original data y t With v t The relationship is expressed as:

[0063]

[0064] The autoregressive moving average model is expressed as:

[0065] y t =a1y t-1 +a2y t-2 +…+a m y t-m +α t +c1α t-1 +c2α t-2 +…c nα t-n

[0066] Among them, t represents the time point, m represents the order of the autoregressive part, and a m represents the autoregressive coefficient, c n represents the moving average coefficient, α t is a normally distributed white noise, and the functional relationship between the output power P1 of the wind turbine and the wind speed V is expressed as:

[0067]

[0068] Among them, V a Indicates the cut-in wind speed of the wind turbine, V b Indicates the rated wind speed of the wind turbine, V c Indicates the cut-out wind speed of the wind turbine, P e Represents the rated power of the wind turbine generator, and a, b, and c are calculated as follows:

[0069]

[0070] Among them, a is the initial slope influence coefficient of the wind turbine output power curve, b is the slope influence coefficient of the power curve when the wind speed increases moderately, and c is the slope influence coefficient of the power curve at the rated wind speed.

[0071] It should also be noted that modeling wind turbine output and photovoltaic output also includes using software to establish a light intensity model to obtain a regional solar radiation intensity model. The output power of the photovoltaic power generation system is positively correlated with the solar radiation intensity. The output model of the photovoltaic power generation system is expressed as:

[0072] P2=rAρK

[0073] Among them, P2 is the photovoltaic power output power, r represents the solar radiation intensity, A represents the photovoltaic array area, ρ represents the conversion efficiency of photovoltaic modules to convert light energy into electrical energy, and K represents the correction factor.

[0074] It should also be noted that the distribution network data collection process, including the integration of distribution network components, structures, environmental operating parameters, historical fault data and reliability evaluation indicators, constitutes a comprehensive data foundation. The collection of these data not only provides necessary input information for subsequent analysis and model building, but also helps to achieve dynamic monitoring and real-time optimization of the distribution network, improve the operating efficiency and reliability of the distribution network, and use an autoregressive moving average model to model and predict wind speed, as well as establish a light intensity model. The present invention provides a method for accurately predicting the output of distributed power sources, which can effectively evaluate the performance of wind power generation and photovoltaic power generation under different environmental conditions, and provide reliable decision support for energy management and scheduling of distribution networks, especially in terms of wind speed and solar radiation. In the case of large intensity fluctuations, an accurate output prediction model can significantly reduce power losses caused by prediction errors, optimize the energy distribution of the distribution network, and enhance the stability of the system. By adopting the light intensity model and autoregressive moving average model established by software, the present invention realizes high-precision modeling of wind turbine and photovoltaic output. These models can not only predict the actual output of distributed power sources, but also evaluate the impact of different weather conditions and environmental changes on the output, thereby providing strong technical support for the operation and maintenance of the distribution network. By accurately modeling and predicting the output of these renewable energy sources, the distribution network's ability to accept renewable energy can be effectively improved, the use of clean energy can be promoted, carbon emissions can be reduced, and support can be provided for the realization of energy transformation and sustainable development goals.

[0075] S2: Use Zeno polyhedron to model the impact of uncertainty on distributed power output and improve the Monte Carlo method to identify key components.

[0076] Furthermore, the modeling of the impact of uncertainty in the output of distributed power generation includes that each node in the distribution network has a maximum acceptable capacity. When a fault occurs, the circuit breaker is activated, the topology of the distribution network changes, and the voltage of the node is raised by the distributed power generation, resulting in an increase in the node voltage. The real-time voltage deviation does not exceed ±7% of the nominal voltage. The uncertainty of the output of distributed power generation is modeled using Zeno polyhedron, which is expressed as:

[0077]

[0078] in, represents the center of the Zeno polyhedron, located in n-dimensional real space, represents the generating vector of the Zeno polyhedron, p is the number of generating vectors, β i is the scaling factor for the i-th generated vector.

[0079] It should be noted that the key component identification includes that the trouble-free operation time and fault repair time of system components obey the exponential distribution. The sampling form of the trouble-free operation time and fault repair time of a single component is expressed as:

[0080]

[0081] Wherein, TTF represents the trouble-free operation time of the component, TTR represents the fault repair time of the component, λ represents the failure rate of the component, δ represents the repair rate of the component, MTTF represents the average failure time of the component, MTTR represents the average repair time of the component, R1 and R2 represent random numbers with a value interval of (0,1), and the key components of the distribution network are identified by the improved Monte Carlo simulation method; the original parameters of the distribution network are read and the data is initialized, and the initial key component set is determined to include all distributed power sources, and a set of random numbers with the same number of elements as the initial key component set is generated, and the value range of the random number is between (0,1). The trouble-free operation time and fault repair time of a single component are sampled based on the random number to calculate the trouble-free operation time of the initial key component. The fault working time and fault repair time are calculated. The distribution network only includes one-dimensional faults of components in the initial key component set, and the components outside the set operate normally. The system state and the corresponding duration are obtained. The power flow after the system fault is calculated, the node voltage of each node after the fault is determined, and it is judged whether the distributed power supply is offline. If the distributed power supply is offline, the overvoltage impact range caused by the fault is determined. If the distributed power supply is not offline, the operating status of each load point is determined; when the system operation time is less than the simulation time, a set of random numbers with the same number of elements as the initial key component set is regenerated; when the system operation time is greater than or equal to the simulation time, the number of component failures and failure time in the simulation period are counted, the system reliability index is calculated, and the key component set is obtained based on the component failure.

[0082] It should also be noted that the access of distributed power sources to the distribution network requires that the distribution network still meets the power quality requirements and needs to have the ability to quickly detect islands and immediately disconnect from the power grid. On the one hand, the access of distributed power sources to the distribution network will increase the node voltage at the access point. For three-phase distribution networks of 20kV and below, the voltage deviation is required to be no more than ±7% of the nominal voltage. Therefore, each node in the distribution network should have a maximum acceptable capacity. On the other hand, when a fault occurs, the circuit breaker is activated, the topology of the distribution network changes, and the increase in node voltage by the distributed power source may cause the voltage of some nodes to increase. The present invention uses Zeno polyhedron to model the output uncertainty of distributed power sources. Taking the wind turbine output model established by the present invention as an example, the autoregressive sliding average model is used according to The annual output model of the wind turbine is obtained in hours. It is assumed that the uncertainty of the hourly output of the wind turbine is ±30% and the rated capacity is 1MW. If at a certain moment, the predicted power of the wind turbine is 60% of the rated capacity, the possible output power is in the range of [0.42, 0.78] MW. In this output environment, it is necessary to ensure that the node voltage of the wind power access point meets the power quality requirements, that is, the voltage deviation cannot exceed ±7%. After the distributed power source is connected to the distribution network, the voltage of some nodes will increase due to the change of the flow. If the distribution network fails at this time and the load is cut off, resulting in an increase in the voltage on other nodes, it may cause the voltage of some nodes to exceed the limit. The requirements for power quality cause partial load shutdown and overvoltage tripping of the distributed power source, affecting the power supply of the distribution network.

[0083] It should also be noted that the access of distributed power sources such as wind power and photovoltaic power generation increases the complexity and uncertainty of the distribution network. The Zeno polyhedron model used in the present invention describes this uncertainty. The Zeno polyhedron model allows the maximum acceptable capacity of each node in the distribution network and its changes to be accurately represented in the n-dimensional real space, which is crucial for evaluating the impact of the access of distributed power sources on the stability of the distribution network. Through precise modeling, the behavior of the distribution network under specific fault conditions can be effectively identified and predicted, especially in terms of voltage changes caused by distributed power sources, thereby improving the operational safety and reliability of the distribution network. Secondly, the identification method of key components further improves the accuracy of the reliability analysis of the distribution network. Through the improved Monte Carlo simulation method, the key components in the distribution network can be more accurately identified. The performance of these components has a decisive influence on the stability and reliability of the entire system, and the system The fault-free operation time and fault repair time of system components are regarded as random variables following exponential distribution, which not only reflects the reality of the fault characteristics of power system components, but also provides an effective means to simulate and predict the behavior of distribution networks when facing various fault conditions. The application of this method, especially in the case of widespread access to distributed power sources, has greatly improved the accuracy and efficiency of distribution network reliability assessment. The use of BPNN as a neural network model for reliability parameter correction demonstrates the application potential of data-driven methods in power system reliability management. By using distribution network data as a training set, not only can the reliability parameters of key components be accurately corrected, but also the distribution network's adaptability to future faults and changes can be improved. The use of the BPNN model, especially the innovative processing in the data normalization and model training stages, further enhances the model's prediction accuracy and generalization ability.

[0084] S3: Select BPNN as the neural network model, use the distribution network data as the training set and normalize it, and perform reliability parameter correction after training the BPNN model.

[0085] Furthermore, the reliability parameter correction includes selecting BPNN as the neural network model, taking the distribution network data as the training set and normalizing it, and correcting the reliability parameters after training the BPNN model; the input vector X of the BPNN is n It includes the initial reliability parameters of key components, operating environment parameters and reliability indicators calculated based on system historical data, and the output vector Y m is the reliability parameter of the key component after correction; before training BPNN, X n The parameters in have different dimensions, including the average power outage frequency index, the average power outage duration index, the expected value of insufficient power, and the ambient temperature and humidity in the system reliability index. Based on the convergence of BPNN training, the dimension differences between different types of variables are eliminated, and the input sample data is normalized and preprocessed, which is expressed as:

[0086]

[0087] Among them, X 0 Represents the normalized BPNN input vector; the forward propagation process of BPNN from input to output is expressed as:

[0088] Y 0 =f L (…f l (…f2(X 0 )

[0089]

[0090] Among them, Y 0 represents the normalized output vector, L represents the total number of layers of BPNN, and Represent the input vector and output vector of the lth layer respectively, f l (.) represents the corrected output function of the lth layer, f L (.) represents the modified output function of the total number of layers of BPNN, s(.) represents the activation function of the neuron, and W l-1 represents the weight between the l-1th layer and the lth layer, b l Represents the bias of the lth layer, selects the rectified linear unit function as the activation function of the hidden layer and the output layer, and calculates the activation function as:

[0091]

[0092] The modification of weight W and bias b during back propagation is expressed as:

[0093]

[0094]

[0095]

[0096]

[0097]

[0098] Among them, E represents the error, N represents the number of samples in the training set, O and O i They represent the target value of neural network training and the actual value of the i-th iteration respectively, η represents the learning rate, is the loss function E relative to the weight W k The gradient of is the loss function E relative to the bias b kThe gradient of the system is used to calculate the reliability index of the historical data as the training target value, W k and b k Respectively represent the weight and bias of the output layer; after the BPNN training results meet the accuracy requirements, the BPNN output vector is denormalized when calculating the revised key component reliability parameters, which is expressed as:

[0099] Y m =Y 0 (max(Y 0 )-min(Y 0 ))+min(Y 0 )

[0100] Output vector Y m The revised reliability parameter results of key components.

[0101] It should be noted that the content of the present invention covers key steps from the collection and normalization of distribution network data to the use of BPNN model for training and the correction of reliability parameters. Through this series of operations, the operating status of the power grid can be effectively monitored and adjusted in real time, thereby improving the reliability and stability of the entire power grid. This method has obvious advantages over the traditional static adjustment method. The present invention allows the system to be dynamically adjusted according to real-time data, which more accurately reflects the actual operation of the power grid. The application of BPNN enables the present invention to process large amounts of complex data. Through the powerful data processing capabilities of neural networks, potential patterns in the data can be identified. and relationship, providing a scientific basis for the correction of reliability parameters, which not only improves the accuracy of correction, but also provides efficient decision-making support for the operation and maintenance of the power grid. The normalization processing step in the present invention solves the impact of data of different dimensions and ensures the effectiveness and accuracy of model training. The application of the back propagation algorithm enables the model to continuously optimize and adjust weights and biases through iterative learning to minimize errors, thereby achieving accurate correction of reliability parameters. Through the inverse normalization processing, the output of the BPNN model is converted into actual reliability parameters, so that the correction results can be directly applied to the operation and maintenance of the power grid, thereby improving the practicability and operability of the invention.

[0102] Example 2

[0103] Reference Figure 2-Figure 5 , which is an embodiment of the present invention, provides a method for correcting reliability parameters of key components of a distribution network with distributed power access. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0104] Firstly, taking the reconstructed IEEE 33-node distribution network as an example, the distribution network data is collected to model the wind turbine output and photovoltaic output; secondly, the influence of the uncertainty of the distributed power output is modeled by using Zeno polyhedron, and the key components are identified by the improved Monte Carlo method; finally, BPNN is selected as the neural network model, and the distribution network data is used as the training set and normalized. After training the BPNN model, the reliability parameters are corrected. The distribution network structure is as follows: Figure 2 As shown, it includes 32 load nodes, 32 transformers, 5 circuit breakers and 3 distributed power sources. The line data and initial reliability parameters of each power component in the feeder are shown in Tables 1 and 2.

[0105] Table 1 Data table of each branch

[0106]

[0107]

[0108] Table 2 Power supply component reliability parameter table

[0109] type Failure rate Repair time(h) line 0.06 5 transformer 0.03 30 breaker 0.01 4

[0110] After using the data for Monte Carlo simulation to find the key components of the system, the initial reliability parameters, operating environment parameters and system reliability index of the components are used as input, and the historical fault data of the distribution network, operating environment parameters and reliability index are used as training sets to train BPNN. The revised reliability parameters are used as output. The comparison between the system reliability index calculated by the reliability parameters before correction, the system reliability index calculated by the revised reliability parameters and the reliability index obtained by historical data statistics is as follows: Figures 3 to 5 As shown, it can be seen that compared with the reliability parameters before correction, the reliability index calculated by the corrected reliability parameters is closer to the reliability index evaluated by the historical records, the model accuracy is higher, and the correction effect is good. Therefore, the present invention can comprehensively consider the influencing factors of the component reliability parameters, correct the reliability parameters of the key components, and thus improve the calculation and prediction accuracy of the system reliability index.

[0111] Example 3

[0112] Reference Figure 6 , which is an embodiment of the present invention, provides a reliability parameter correction system for key components of a distribution network with distributed power supply access, including a data processing module, a component identification module, and a parameter correction module.

[0113] The data processing module is used to collect distribution network data and model wind turbine output and photovoltaic output; the component identification module is used to use Zeno polyhedron to model the impact of uncertainty on distributed power output and improve the Monte Carlo method to identify key components; the parameter correction module is used to select BPNN as the neural network model, use the distribution network data as the training set and normalize it, and perform reliability parameter correction after training the BPNN model.

[0114] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, 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. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of 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, etc. Various media that can store program codes.

[0115] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0116] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0117] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0118] 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 preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for correcting reliability parameters of key components of a distribution network connected to a distributed power source, characterized in that: include: Collect distribution network data and model wind turbine and photovoltaic output; Zeno polyhedron is used to model the influence of uncertainty on the output of distributed power generation, and the Monte Carlo method is improved to identify key components; BPNN is selected as the neural network model, the distribution network data is used as the training set and normalized, and the reliability parameters are corrected after the BPNN model is trained.

2. The method for correcting reliability parameters of key components of a distribution network for access to distributed power sources according to claim 1, characterized in that: The distribution network data includes distribution network components, distribution network structure, distribution network environmental operating parameters, historical fault data, and reliability evaluation indicators.

3. The method for correcting reliability parameters of key components of a distribution network for access to distributed power sources as claimed in claim 2, characterized in that: The modeling of wind turbine output and photovoltaic output includes using an autoregressive moving average model to model and predict wind speed, collecting regional wind speed data, and based on the original wind speed data v t Calculate the average wind speed μ and the standard deviation σ of the wind speed distribution, and t After standardization, the original data y t With v t The relationship is expressed as: The autoregressive moving average model is expressed as: y t =a1y t-1 +a2y t-2 +…+a m y t-m +a t +c1a t-1 +c2a t-2 +…c n a t-n Among them, t represents the time point, m represents the order of the autoregressive part, and a m represents the autoregressive coefficient, c n represents the moving average coefficient, α t is a normally distributed white noise, and the functional relationship between the output power P1 of the wind turbine and the wind speed V is expressed as: Among them, V a Indicates the cut-in wind speed of the wind turbine, V b Indicates the rated wind speed of the wind turbine, V c Indicates the cut-out wind speed of the wind turbine, P e Represents the rated power of the wind turbine generator, and a, b, and c are calculated as follows: Among them, a is the initial slope influence coefficient of the wind turbine output power curve, b is the slope influence coefficient of the power curve when the wind speed increases moderately, and c is the slope influence coefficient of the power curve at the rated wind speed.

4. The method for correcting reliability parameters of key components of a distribution network connected to a distributed power source according to claim 3, characterized in that: The modeling of wind turbine output and photovoltaic output also includes using software to establish a light intensity model to obtain a regional solar radiation intensity model. The output power of the photovoltaic power generation system is positively correlated with the solar radiation intensity. The output model of the photovoltaic power generation system is expressed as: P2=rAρK Among them, P2 is the photovoltaic power output power, r represents the solar radiation intensity, A represents the photovoltaic array area, ρ represents the conversion efficiency of photovoltaic modules to convert light energy into electrical energy, and K represents the correction factor.

5. The method for correcting reliability parameters of key components of a distribution network for access to distributed power sources as claimed in claim 4, characterized in that: The modeling of the impact of uncertainty on the output of distributed power sources includes that each node in the distribution network has a maximum acceptable capacity. When a fault occurs, the circuit breaker is activated, the topology of the distribution network changes, and the node voltage is raised by the distributed power source, resulting in an increase in the node voltage. The real-time voltage deviation does not exceed ±7% of the nominal voltage. The uncertainty of the output of distributed power sources is modeled using Zeno polyhedron, which is expressed as: in, represents the center of the Zeno polyhedron, located in n-dimensional real space, represents the generating vector of the Zeno polyhedron, p is the number of generating vectors, β i is the scaling factor for the i-th generated vector.

6. The method for correcting reliability parameters of key components of a distribution network connected to a distributed power source according to claim 5, characterized in that: The key component identification includes that the trouble-free operation time and the fault repair time of the system components are all subject to exponential distribution, and the sampling form of the trouble-free operation time and the fault repair time of a single component is expressed as: Among them, TTF represents the trouble-free operation time of the component, TTR represents the fault repair time of the component, λ represents the failure rate of the component, δ represents the repair rate of the component, MTTF represents the mean failure time of the component, MTTR represents the mean repair time of the component, R1 and R2 represent random numbers with a value interval of (0,1), and the key components of the distribution network are identified by the improved Monte Carlo simulation method; Read the original parameters of the distribution network and initialize the data, determine that the initial key component set includes all distributed power sources, generate a set of random numbers with the same number of elements as the initial key component set, and the value range of the random numbers is between (0,1). The fault-free working time and fault repair time of the initial key components are calculated based on the random numbers in the sampling form of the fault-free operating time and fault repair time of a single component. The distribution network only includes one-dimensional faults of the components in the initial key component set, and the components outside the set operate normally. The system state and the corresponding duration are obtained, the power flow after the system fault is calculated, the node voltage of each node after the fault is determined, and it is judged whether the distributed power source is offline. If the distributed power source is offline, the overvoltage impact range caused by the fault is determined. If the distributed power source is not offline, the operating state of each load point is determined; When the system running time is less than the simulation time, a set of random numbers with the same number of elements as the initial key element set is regenerated; When the system operation time is greater than or equal to the simulation time, the number of component failures and failure time during the simulation period are counted, the system reliability index is calculated, and the key component set is obtained based on component failures.

7. The method for correcting reliability parameters of key components of a distribution network connected to a distributed power source according to claim 6, characterized in that: The reliability parameter correction includes selecting BPNN as a neural network model, using the distribution network data as a training set and normalizing it, and correcting the reliability parameters after training the BPNN model; BPNN input vector X n It includes the initial reliability parameters of key components, operating environment parameters and reliability indicators calculated based on system historical data, and the output vector Y m Corrected reliability parameters for key components; Before training BPNN, X n The parameters in have different dimensions, including the average power outage frequency index, the average power outage duration index, the expected value of insufficient power, and the ambient temperature and humidity in the system reliability index. Based on the convergence of BPNN training, the dimension differences between different types of variables are eliminated, and the input sample data is normalized and preprocessed, which is expressed as: Among them, X 0 represents the normalized BPNN input vector; The forward propagation process of BPNN from input to output is expressed as: Y 0 =f L (…f l (…f2(X 0 ) Among them, Y 0 represents the normalized output vector, L represents the total number of layers of BPNN, and Represent the input vector and output vector of the lth layer respectively, f l (.) represents the corrected output function of the lth layer, f L (.) represents the modified output function of the total number of layers of BPNN, s(.) represents the activation function of the neuron, and W l-1 represents the weight between the l-1th layer and the lth layer, b l Represents the bias of the lth layer, selects the rectified linear unit function as the activation function of the hidden layer and the output layer, and calculates the activation function as: The modification of weight W and bias b during back propagation is expressed as: Among them, E represents the error, N represents the number of samples in the training set, O and O i They represent the target value of neural network training and the actual value of the i-th iteration respectively, η represents the learning rate, is the loss function E relative to the weight W k The gradient of is the loss function E relative to the bias b k The gradient of the system is used to calculate the reliability index of the historical data as the training target value, W k and b k Represent the weight and bias of the output layer respectively; After the BPNN training results meet the accuracy requirements, the BPNN output vector is denormalized when calculating the revised key component reliability parameters, which is expressed as: AND m =And 0 (max(Y 0 )-min(Y 0 ))+min(Y 0 ) Output vector Y m The revised reliability parameter results of key components.

8. A system using the method for correcting reliability parameters of key components of a distribution network in which distributed power sources are connected as claimed in any one of claims 1 to 7, characterized in that: It includes data processing module, component identification module and parameter correction module; The data processing module is used to collect distribution network data and model wind turbine output and photovoltaic output; The component identification module is used to model the influence of uncertainty on the output of distributed power sources using Zeno polyhedrons and to identify key components using an improved Monte Carlo method; The parameter correction module is used to select BPNN as the neural network model, use the distribution network data as a training set and normalize it, and perform reliability parameter correction after training the BPNN model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for correcting reliability parameters of key components of a distribution network to which distributed power sources are connected as described in any one of claims 1 to 7 are implemented.

Images