Power distribution network fault active preprocessing method and device based on distributed energy storage

Abstract

The invention discloses a power distribution network fault active preprocessing method and device based on distributed energy storage, and the method comprises the steps: obtaining the operation parameter data of a power distribution network based on the data monitoring of the power distribution network, constructing a power distribution network state time sequence prediction model based on a long short-term memory neural network, and predicting the future operation parameter data of the power distribution network through the model; when future operation parameter data of the power distribution network are predicted through the model, an attention mechanism in the model performs weighted combination on a hidden state sequence to highlight information of key time steps, and whether key parameters of the power distribution network are out of limit or not is judged through obtained prediction data; constructing a cooperative control model based on distributed energy storage, selecting a corresponding distributed energy storage operation strategy according to whether key parameters of the power distribution network are out of limit or not, and performing active preprocessing on faults of the power distribution network; the power distribution network fault active preprocessing strategy based on distributed energy storage has important significance for improving the safe and stable operation level and the energy utilization efficiency of the power distribution network.

Description

Technical Field

[0001] The present invention belongs to the field of distribution network fault early warning methods, and in particular relates to a distribution network fault active preprocessing method and device based on distributed energy storage. Background Art

[0002] With the rapid development of the power system, the safe and stable operation of the distribution network, as a key link connecting power sources and users, is of great significance to ensuring power supply and promoting economic and social development. However, with the large-scale access of distributed renewable energy and the increasing diversification of load demand, the distribution network faces a more complex operating environment, and the probability and risk of failures also increase. Traditional distribution network fault handling strategies mostly adopt a passive response approach, that is, fault location and recovery after the fault occurs. This approach is not only time-consuming, but also easy to cause a great impact on the power grid and users. Therefore, how to achieve active pre-processing of distribution network faults and improve the efficiency and accuracy of fault handling has become an important issue in the current operation and management of distribution networks.

[0003] As a new type of energy storage method, distributed energy storage technology has the advantages of strong flexibility, fast response speed, and easy deployment, and has been widely used in power systems. By connecting the distributed energy storage system to the distribution network, real-time monitoring and rapid adjustment of the power grid status can be achieved, providing a new technical means for active pre-processing of distribution network faults. Distributed energy storage systems can not only serve as a backup power source to provide power support when a fault occurs, but also optimize control strategies to adjust and balance the grid load and reduce the probability and risk of faults. Therefore, in-depth research on active pre-processing strategies for distribution network faults based on distributed energy storage is of great significance to improving the safe and stable operation level and energy utilization efficiency of the distribution network. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art, based on the distribution network state prediction, and to conduct in-depth research on the active preprocessing strategy of distribution network faults based on distributed energy storage, which is of great significance for improving the safe and stable operation level and energy utilization efficiency of the distribution network.

[0005] The present invention adopts the following technical solutions:

[0006] The method for active preprocessing of distribution network faults based on distributed energy storage comprises the following steps:

[0007] Based on the data monitoring of the distribution network, the distribution network operation parameter data is obtained, and a distribution network state time series prediction model based on long short-term memory neural network is constructed, and the future operation parameter data of the distribution network is predicted through the model;

[0008] When predicting the future operating parameter data of the distribution network through the model, the attention mechanism in the model performs weighted combination on the hidden state sequence to highlight the information of the key time step, and judges whether the key parameters of the distribution network exceed the limit through the obtained prediction data;

[0009] A collaborative control model based on distributed energy storage is constructed. According to whether the key parameters of the distribution network exceed the limit, the corresponding distributed energy storage operation strategy is selected to actively pre-process the distribution network faults.

[0010] Furthermore, the distribution network operating parameter data are index data that affect the operating state of the distribution network, including the voltage and phase angle of each node, the current of each branch, the power factor, the active power and the reactive power.

[0011] Furthermore, the attention mechanism in the model performs weighted combination on the hidden state sequence to highlight the information of the key time step as follows:

[0012] After the long short-term memory neural network processes the time series, it obtains the hidden state sequence:

[0013]

[0014] Attention score calculation: For each hidden state h t , calculate the attention score:

[0015]

[0016] Among them, W a and b a are the trainable parameters of the attention layer;

[0017] Weighted sum of hidden states by attention weights:

[0018]

[0019] Combine the vector c with the last hidden state h T After combining, input the fully connected layer for prediction:

[0020]

[0021] W c and W o is the weight of the fully connected layer, b c and b o is the bias term, is the final prediction of the model.

[0022] Furthermore, the coordinated control model based on distributed energy storage includes a first objective function and a second objective function, wherein:

[0023] The first objective function is:

[0024]

[0025] The second objective function is:

[0026]

[0027] It also includes constraints, which include distribution network constraints and energy storage system constraints.

[0028] Furthermore, according to whether the key parameters of the distribution network exceed the limit, the method of selecting the corresponding distributed energy storage operation strategy is as follows:

[0029] If both the voltage and reactive power exceed the limit, the first objective function of the coordinated control model based on distributed energy storage is used to actively pre-process the distribution network fault;

[0030] If both the frequency and active power exceed the limit, the second objective function of the coordinated control model based on distributed energy storage is used to actively pre-process the distribution network fault.

[0031] Furthermore, the distribution network constraint is:

[0032] In a distribution network, the power of the system is defined by the voltage, impedance, and current at the node. For each node, the voltage V i and power P i and Q i There are the following relationships between them:

[0033]

[0034] Among them, V i and V j is the voltage amplitude at node i and node j; G ij , B ij are the real and imaginary parts of the admittance matrix between node i and node j respectively; θ i is the voltage phase angle between node i and node j; P i is the active power of the ith node; Q i is the reactive power of the ith node;

[0035] In the distribution network, voltage is the main control parameter of each node; usually, the magnitude and phase of voltage are subject to certain constraints:

[0036]

[0037] and are the minimum and maximum voltage limits of the ith node voltage respectively;

[0038] The active power P of the kth branch (from node i to node j) ij The transmission capacity is limited by the maximum power transmission capacity of the branch The formula is:

[0039]

[0040] Similarly, the reactive power transfer Q of the kth branch ij There is also a minimum and maximum constraint:

[0041]

[0042] The energy storage system constraints are:

[0043] The charging power of the energy storage system is usually limited by the maximum charging power and the minimum charging power. The charging power of the battery cannot exceed the maximum value allowed by the equipment, and cannot be less than a minimum value allowed. The constraints of its charging and discharging power are:

[0044]

[0045] is the charging power of energy storage unit k at time t (usually a positive value indicates charging); is the maximum charging power of energy storage unit k; is the discharge power of energy storage unit k at time t (usually a positive value indicates discharge); is the maximum discharge power of energy storage unit k;

[0046] The charging and discharging power of the energy storage system is constrained by the battery capacity; the charging and discharging power of the battery depends not only on the maximum power limit, but also on the current state of the battery (such as the battery's state of charge SOC); the relationship between battery capacity and state of charge can be expressed as:

[0047]

[0048] SOC k (t) is the state of charge of energy storage unit k at time t; and are the minimum and maximum states of charge of energy storage unit k, respectively;

[0049] The relationship between the state of charge and the charge and discharge power is usually modeled through the charge and discharge efficiency and power curve. For example, changes in the charging power and discharge power will affect the SOC of the energy storage unit, and the formula is:

[0050]

[0051] C kis the battery capacity of energy storage unit k (usually in Ah); and are the charging and discharging efficiencies of energy storage unit k (usually less than 1); Δt is the time step length (usually hours or minutes).

[0052] The active pre-processing device for distribution network faults based on distributed energy storage includes:

[0053] A distribution network future operation parameter data acquisition module is used to acquire distribution network operation parameter data based on distribution network data monitoring, build a distribution network state time series prediction model based on long short-term memory neural network, and predict the distribution network future operation parameter data through the model;

[0054] A distribution network key parameter over-limit judgment module is used to predict the future operating parameter data of the distribution network through the model. The attention mechanism in the model performs weighted combination on the hidden state sequence to highlight the information of the key time step, and judge whether the key parameters of the distribution network are over-limit through the obtained prediction data;

[0055] The distribution network fault preprocessing module is used to build a collaborative control model based on distributed energy storage. According to whether the key parameters of the distribution network are out of limit, the corresponding distributed energy storage operation strategy is selected to actively preprocess the distribution network fault.

[0056] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned active preprocessing method for distribution network faults based on distributed energy storage.

[0057] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for active preprocessing of distribution network faults based on distributed energy storage as described above is implemented.

[0058] The beneficial technical effects of the present invention are:

[0059] 1. Distribution network operation data usually contains complex nonlinear relationships (such as load changes, fault modes, etc.). The LSTM combined with the attention mechanism of the present invention can better capture these relationships and provide accurate predictions; by dynamically adjusting the focus, the model can better cope with time series of different modes or external disturbances (such as seasonal load changes, short-term abnormal fluctuations, etc.). The emphasis of the attention mechanism on important time steps makes the model more suitable for real-time prediction scenarios (such as distribution network status monitoring), which provides a reliable reference for power grid operation and dispatching;

[0060] (2) Distributed energy storage can detect potential problems and take measures (such as adjusting power output, peak shaving and valley filling) before a fault occurs through real-time monitoring and rapid response mechanisms, thereby preventing the fault from spreading further, reducing power outages caused by equipment failures, and improving the reliability of power system operation. Distributed energy storage can actively adjust the voltage level and frequency of nodes, especially in fault areas or weak links, to ensure that the power quality is not affected by short-term disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 Block diagram of a method for active preprocessing of distribution network faults based on distributed energy storage provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following is a further clear and complete description of a method and device for clearing a power capacity market taking into account flexibility provided by the present invention in conjunction with the accompanying drawings:

[0063] Example 1

[0064] like Figure 1 As shown, the present embodiment provides a method for active preprocessing of distribution network faults based on distributed energy storage, the method comprising the following steps:

[0065] Based on the data monitoring of the distribution network, the distribution network operation parameter data is obtained, and a distribution network state time series prediction model based on long short-term memory neural network is constructed, and the future operation parameter data of the distribution network is predicted through the model; specifically:

[0066] Collect multi-source data of the distribution network, including historical load data, equipment operation data, meteorological data (such as temperature and humidity), topology information, and historical fault records. Extract key indicators that affect the operation status of the distribution network, namely, the voltage and phase angle of each node, the current of each branch, power factor, active power, and reactive power.

[0067] Assume that the input data of the distribution network is a historical time series X = {x1, x2, ..., x T}, where x t Represents the characteristic vector at time t, such as voltage and phase angle;

[0068] The goal is to predict the running state Y = {y T+1 ,y T+2 ,...,y T+K}, where y T+k is the target output (such as voltage and phase angle state);

[0069] Normalize the data to prevent the numerical ranges of different features from affecting model training:

[0070]

[0071] The LSTM unit consists of an input gate, a forget gate, and an output gate to control the flow of information; the core formula is as follows:

[0072] Forget gate: decides which past information to forget:

[0073] f t =σ(W f ·[h t-1 ,x t ]+b f ) (2)

[0074] Input gate: decides which new input information to keep:

[0075]

[0076] Status update: Update the memory cell status of the current cell:

[0077]

[0078] Output gate: determines what information to output:

[0079]

[0080] Among them, σ is the Sigmoid activation function, tanh is the hyperbolic tangent function, and W f ,W i ,W C ,W o and b f ,b i ,b C ,b o .Model weights and biases; h t is the hidden state of LSTM (feature representation of the encoding sequence), C t is the cell state of LSTM.

[0081] When predicting the future operating parameter data of the distribution network through the model, the attention mechanism in the model performs weighted combination on the hidden state sequence to highlight the information of the key time step, and judges whether the key parameters of the distribution network exceed the limit through the obtained prediction data;

[0082] Among them, after the long short-term memory neural network processes the time series, it obtains the hidden state sequence:

[0083]

[0084] Attention score calculation: For each hidden state h t , calculate the attention score:

[0085]

[0086] Among them, W a and b a are the trainable parameters of the attention layer;

[0087] Weighted sum of hidden states by attention weights:

[0088]

[0089] Combine the vector c with the last hidden state h T After combining, input the fully connected layer for prediction:

[0090]

[0091] W c and W o is the weight of the fully connected layer, b c and b o is the bias term, is the final prediction of the model.

[0092] Construct a coordinated control model based on distributed energy storage, select the corresponding distributed energy storage operation strategy according to whether the key parameters of the distribution network exceed the limit, and actively pre-process the distribution network fault; wherein, the coordinated control model based on distributed energy storage includes a first objective function and a second objective function, wherein,

[0093] The first objective function is:

[0094]

[0095] The second objective function is:

[0096]

[0097] It also includes constraints, which include distribution network constraints and energy storage system constraints.

[0098] According to whether the key parameters of the distribution network exceed the limit, the method of selecting the corresponding distributed energy storage operation strategy is as follows:

[0099] If both the voltage and reactive power exceed the limit, the first objective function of the coordinated control model based on distributed energy storage is used to actively pre-process the distribution network fault;

[0100] If both the frequency and active power exceed the limit, the second objective function of the coordinated control model based on distributed energy storage is used to actively pre-process the distribution network fault.

[0101] The distribution network constraints are:

[0102] In a distribution network, the power of the system is defined by the voltage, impedance, and current at the node. For each node, the voltage V i and power P i and Q i There are the following relationships between them:

[0103]

[0104] Among them, V i and V j is the voltage amplitude at node i and node j; G ij , B ij are the real and imaginary parts of the admittance matrix between node i and node j respectively; θ i is the voltage phase angle between node i and node j; P i is the active power of the ith node; Q i is the reactive power of the ith node;

[0105] In the distribution network, voltage is the main control parameter of each node; usually, the magnitude and phase of voltage are subject to certain constraints:

[0106]

[0107] and are the minimum and maximum voltage limits of the ith node voltage respectively;

[0108] The active power P of the kth branch (from node i to node j) ij The transmission capacity is limited by the maximum power transmission capacity of the branch The formula is:

[0109]

[0110] Similarly, the reactive power transfer Q of the kth branch ij There is also a minimum and maximum constraint:

[0111]

[0112] The energy storage system constraints are:

[0113] The charging power of the energy storage system is usually limited by the maximum charging power and the minimum charging power. The charging power of the battery cannot exceed the maximum value allowed by the equipment, and cannot be less than a minimum value allowed. The constraints of its charging and discharging power are:

[0114]

[0115] is the charging power of energy storage unit k at time t (usually a positive value indicates charging); is the maximum charging power of energy storage unit k; is the discharge power of energy storage unit k at time t (usually a positive value indicates discharge); is the maximum discharge power of energy storage unit k;

[0116] The charging and discharging power of the energy storage system is constrained by the battery capacity; the charging and discharging power of the battery depends not only on the maximum power limit, but also on the current state of the battery (such as the battery's state of charge SOC); the relationship between battery capacity and state of charge can be expressed as:

[0117]

[0118] SOC k (t) is the state of charge of energy storage unit k at time t; and are the minimum and maximum states of charge of energy storage unit k, respectively;

[0119] The relationship between the state of charge and the charge and discharge power is usually modeled through the charge and discharge efficiency and power curve. For example, changes in the charging power and discharge power will affect the SOC of the energy storage unit, and the formula is:

[0120]

[0121] C k is the battery capacity of energy storage unit k (usually in Ah); and are the charging and discharging efficiencies of energy storage unit k (usually less than 1); Δt is the time step length (usually hours or minutes).

[0122] Example 2

[0123] This embodiment provides a distribution network fault active pre-processing device based on distributed energy storage, including:

[0124] A distribution network future operation parameter data acquisition module is used to acquire distribution network operation parameter data based on distribution network data monitoring, build a distribution network state time series prediction model based on long short-term memory neural network, and predict the distribution network future operation parameter data through the model;

[0125] A distribution network key parameter over-limit judgment module is used to predict the future operating parameter data of the distribution network through the model. The attention mechanism in the model performs weighted combination on the hidden state sequence to highlight the information of the key time step, and judge whether the key parameters of the distribution network are over-limit through the obtained prediction data;

[0126] The distribution network fault preprocessing module is used to build a collaborative control model based on distributed energy storage. According to whether the key parameters of the distribution network are out of limit, the corresponding distributed energy storage operation strategy is selected to actively preprocess the distribution network fault.

[0127] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned active preprocessing method for distribution network faults based on distributed energy storage.

[0128] Furthermore, the present invention adopts the following technical solutions:

[0129] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for active preprocessing of distribution network faults based on distributed energy storage as described above is implemented.

[0130] Through the description of the above embodiments, it can be clearly understood by those skilled in the art that the facilities of the present invention can be implemented by means of software plus the necessary general hardware platform. The embodiments of the present invention can be implemented using an existing processor, or by a dedicated processor used for this purpose or other purposes for an appropriate system, or by a hard-wired system. The embodiments of the present invention also include a non-transitory computer-readable storage medium, which includes a machine-readable medium for carrying or having a machine-executable instruction or data structure stored thereon; such a machine-readable medium can be any available medium that can be accessed by a general or special-purpose computer or other machine with a processor. For example, such a machine-readable medium can include RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or any other medium that can be used to carry or store the required program code in the form of machine-executable instructions or data structures, and can be accessed by a general or special-purpose computer or other machine with a processor. When information is transmitted or provided to a machine via a network or other communication connection (hard wiring, or wireless, or a combination of hard wiring and wireless), the connection is also considered a machine-readable medium.

[0131] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. An active preprocessing method for distribution network faults based on distributed energy storage, characterized in that: The method comprises the following steps: Based on the data monitoring of the distribution network, the distribution network operation parameter data is obtained, and a distribution network state time series prediction model based on long short-term memory neural network is constructed, and the future operation parameter data of the distribution network is predicted through the model; When predicting the future operating parameter data of the distribution network through the model, the attention mechanism in the model performs weighted combination on the hidden state sequence to highlight the information of the key time step, and judges whether the key parameters of the distribution network exceed the limit through the obtained prediction data; A collaborative control model based on distributed energy storage is constructed. According to whether the key parameters of the distribution network exceed the limit, the corresponding distributed energy storage operation strategy is selected to actively pre-process the distribution network faults.

2. The method for active preprocessing of distribution network faults based on distributed energy storage according to claim 1 is characterized in that: The distribution network operation parameter data are index data that affect the operation status of the distribution network, including the voltage and phase angle of each node, the current of each branch, the power factor, the active power and the reactive power.

3. The method for active preprocessing of distribution network faults based on distributed energy storage according to claim 1 is characterized in that: The attention mechanism in the model performs weighted combination of hidden state sequences to highlight the information of key time steps as follows: After the long short-term memory neural network processes the time series, it obtains the hidden state sequence: Attention score calculation: For each hidden state h t , calculate the attention score: Among them, W a and b a are the trainable parameters of the attention layer; Weighted sum of hidden states by attention weights: Combine the vector c with the last hidden state h T After combining, input the fully connected layer for prediction: W c and W o is the weight of the fully connected layer, b c and b o is the bias term, is the final prediction of the model.

4. The method for active preprocessing of distribution network faults based on distributed energy storage according to claim 1 is characterized in that: The coordinated control model based on distributed energy storage includes a first objective function and a second objective function, where: The first objective function is: The second objective function is: It also includes constraints, which include distribution network constraints and energy storage system constraints.

5. The method for active preprocessing of distribution network faults based on distributed energy storage according to claim 4 is characterized in that: According to whether the key parameters of the distribution network exceed the limit, the method of selecting the corresponding distributed energy storage operation strategy is as follows: If both the voltage and reactive power exceed the limit, the first objective function of the coordinated control model based on distributed energy storage is used to actively pre-process the distribution network fault; If both the frequency and active power exceed the limit, the second objective function of the coordinated control model based on distributed energy storage is used to actively pre-process the distribution network fault.

6. The method for active preprocessing of distribution network faults based on distributed energy storage according to claim 1 is characterized in that: The distribution network constraints are: In the distribution network, for each node, the voltage V i and power P i and Q i There are the following relationships between them: Among them, V i and V j is the voltage amplitude at node i and node j; G ij , B ij are the real and imaginary parts of the admittance matrix between node i and node j respectively; θ i is the voltage phase angle between node i and node j; P i is the active power of the ith node; Q i is the reactive power of the ith node; In the distribution network, the voltage constraint is: and are the minimum and maximum voltage limits of the ith node voltage respectively; The active power P of the kth branch (from node i to node j) ij The transmission capacity is limited by the maximum power transmission capacity of the branch The formula is: Similarly, the reactive power transfer Q of the kth branch ij Constraints: The energy storage system constraints are: The constraints of the charging and discharging power of the energy storage system are: is the charging power of energy storage unit k at time t; is the maximum charging power of energy storage unit k; is the discharge power of energy storage unit k at time t; is the maximum discharge power of energy storage unit k; The charging and discharging power of the energy storage system is constrained by the battery capacity; the relationship between battery capacity and state of charge can be expressed as: SOC k (t) is the state of charge of energy storage unit k at time t; and are the minimum and maximum states of charge of energy storage unit k, respectively; The relationship between state of charge and charge and discharge power is usually modeled through charge and discharge efficiency and power curves.

7. An active pre-processing device for distribution network faults based on distributed energy storage, characterized in that: include: A distribution network future operation parameter data acquisition module is used to acquire distribution network operation parameter data based on distribution network data monitoring, build a distribution network state time series prediction model based on long short-term memory neural network, and predict the distribution network future operation parameter data through the model; A distribution network key parameter over-limit judgment module is used to predict the future operating parameter data of the distribution network through the model. The attention mechanism in the model performs weighted combination on the hidden state sequence to highlight the information of the key time step, and judge whether the key parameters of the distribution network are over-limit through the obtained prediction data; The distribution network fault preprocessing module is used to build a collaborative control model based on distributed energy storage. According to whether the key parameters of the distribution network are out of limit, the corresponding distributed energy storage operation strategy is selected to actively preprocess the distribution network fault.

8. 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 active preprocessing of distribution network faults based on distributed energy storage as described in any one of claims 1 to 6 is implemented.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for active preprocessing of distribution network faults based on distributed energy storage is implemented as described in any one of claims 1 to 6.

Images