Centralized compensation method for communication time delay of power distribution network

Through the combined prediction and roulette algorithm of LSTM and ARMA models, the inaccurate power regulation caused by the communication delay of the distribution network and the output fluctuation of new energy is solved, and accurate prediction and delay compensation of new energy output power are achieved. , ensuring the safe and efficient operation of the distribution network.

CN118972253AInactive Publication Date: 2024-11-15陈苏程
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202411113415.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The delay of the distribution network communication network and the fluctuation of new energy output lead to inaccurate power regulation and even inability to regulate, affecting the safety of the power grid.

Method used

The LSTM network and ARMA model are used to predict the power generation power of new energy power generation units in combination, and power regulation instructions are formulated based on historical delay data and selection probability, and communication delay compensation is performed through roulette algorithm.

Benefits of technology

It realizes accurate prediction of new energy output power and effective compensation of communication delay, ensuring the safe and efficient operation of the distribution network.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118972253A_ABST
    Figure CN118972253A_ABST
Patent Text Reader

Abstract

The invention discloses a centralized compensation method for communication time delay of a power distribution network, which belongs to the field of communication time delay compensation of the power distribution network, and comprises the following steps: obtaining the maximum communication time delay of all nodes based on the historical time delay data of each power node communication network of the power distribution network; performing combined prediction on the generated power of the new energy power generation unit by using an LSTM network and an ARMA model to obtain a new energy output prediction value; the maximum communication time delay is equally divided into a plurality of time periods, and the selection probability of each time period is obtained based on historical time delay data; and formulating a power regulation and control instruction based on the selection probability of each time period and the new energy output predicted value. According to the method, the estimation of the communication network time delay selection interval can be realized, and the problem that the new energy power caused by the power distribution network communication time delay is difficult to accurately control is solved by utilizing predictive compensation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network communication delay compensation, and in particular relates to a centralized compensation method for distribution network communication delay. Background Art

[0002] In recent years, with the massive access of new energy sources, the development of distribution networks has brought many challenges. The communication network of traditional distribution networks has problems such as delay and packet loss. Due to the volatility and randomness of the output of new energy sources, the power control instructions of the distribution network are affected by the communication network delay and packet loss during the issuance process, resulting in inaccurate power control or even inability to control. In serious cases, it even endangers the safety of the power grid. Therefore, the problem of communication network delay should not be underestimated. The existing power control methods generally do not consider the impact of communication delay when formulating power control instructions. The present invention aims at the problem of inaccurate power control of new energy sources caused by the delay of the distribution network communication network. It is based on a centralized communication delay compensation mechanism and a combined prediction of the power output of new energy sources to cope with the impact of communication network delay, thereby ensuring the safe and efficient operation of the distribution network. Summary of the invention

[0003] In order to solve the above technical problems, the present invention proposes a centralized compensation method for communication delay in a distribution network to solve the above problems existing in the prior art.

[0004] To achieve the above object, the present invention provides a centralized compensation method for communication delay of a distribution network, comprising:

[0005] Based on the historical delay data of the communication network of each power node in the distribution network, the maximum communication delay of all nodes is obtained;

[0006] The LSTM network and ARMA model are used to combine and predict the power generation of the renewable energy power generation unit to obtain the predicted value of renewable energy output;

[0007] Divide the maximum communication delay into several time periods, and obtain the selection probability of each time period based on the historical delay data;

[0008] A power control instruction is formulated based on the selection probability of each time period and the predicted value of the new energy output.

[0009] Optionally, the process of using the LSTM network and the ARMA algorithm to perform combined prediction on the power generation of the new energy power generation unit includes:

[0010] Get 24-hour historical data of new energy power generation units;

[0011] Standardizing the 24-hour historical data to obtain normalized data;

[0012] Constructing an LSTM network, and inputting the normalized data into the LSTM network to obtain a first prediction value;

[0013] Constructing an ARMA model, and inputting the normalized data into the ARMA model to obtain a second predicted value;

[0014] The first prediction value and the second prediction value are combined and weighted averaged to obtain a new energy output prediction value.

[0015] Optionally, the acquired 24-hour historical data of the new energy power generation unit includes: output historical data, meteorological historical data and date type data.

[0016] Optionally, the process of constructing the ARMA model includes: superimposing and combining the autoregressive model and the moving average model to obtain the ARMA model; wherein the expression of the ARMA model is:

[0017]

[0018] In the formula, y t and t-i They refer to the current value and the value at the previous i moments respectively; μ is a constant term, p refers to the order of the autoregressive model, and γ i refers to the autocorrelation coefficient, ε t and ε t-i They refer to the error at the current moment and the error at the previous i moments, respectively. q refers to the order of the moving average model. θ i Refers to the partial autocorrelation coefficient.

[0019] Optionally, the expression for calculating the predicted value of new energy output is:

[0020]

[0021] In the formula, y i It refers to the predicted value of the LSTM network at the current moment, z t Represents the predicted value of new energy output.

[0022] Optionally, the expression for the selection probability of each time period is:

[0023]

[0024] Where P i The communication delay in the time period The probability of The communication delay in the time period The number of times, N τ is the total amount of communication delay data.

[0025] Optionally, the process of formulating a power control instruction and issuing a communication delay compensation amount based on the selection probability of each time period and the predicted value of the new energy output includes: selecting the communication delay compensation amount of each node using a roulette algorithm based on the selection probability of each time period and the predicted value of the new energy output;

[0026] Among them, the expression of communication delay compensation is:

[0027]

[0028] In the formula, is the new energy output adjustment instruction of the jth node, is the first The predicted value at time.

[0029] Compared with the prior art, the present invention has the following advantages and technical effects:

[0030] The present invention can effectively overcome the problem of difficult to accurately control the output power caused by the communication network delay of the distribution network and the randomness and volatility of the output of renewable energy. It realizes accurate prediction of renewable energy output through combined prediction based on LSTM and ARMA algorithms, and estimates the maximum communication delay of each node by using roulette algorithm, which effectively compensates for the power deviation caused by the communication delay and ensures the safe and efficient operation of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0032] Figure 1 This is a flow chart of a centralized compensation method for communication delay in a distribution network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0034] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0035] Embodiment 1

[0036] The existing power distribution communication network is difficult to achieve accurate optimization and scheduling due to the influence of communication delay and volatility of new energy output. The purpose of the present invention is to provide a centralized compensation method for distribution network communication delay to solve the problem that the existing distribution network communication network delay affects the accurate scheduling of new energy power, thereby ensuring the safe and efficient operation of the distribution network. The present invention obtains the maximum communication delay of all nodes based on the historical delay data of the communication network of each power node in the distribution network; uses LSTM network and ARMA model to combine and predict the power generation of new energy power generation units to obtain the predicted value of new energy output; divides the maximum communication delay into several time periods, and obtains the selection probability of each time period based on the historical delay data; formulates power control instructions based on the selection probability of each time period and the predicted value of new energy output.

[0037] like Figure 1 As shown, this embodiment provides a centralized compensation method for distribution network communication delay, including the following steps:

[0038] Step 1: Collect the delay data of the communication network of each power node in the distribution network, establish a database, and obtain the maximum communication delay of all nodes through the historical data of the communication network delay;

[0039] Step 2: Based on the maximum communication delay of all nodes obtained in step 1, a combined prediction of the power generation of the new energy power generation units in the distribution system is performed based on the LSTM network and the ARMA algorithm.

[0040] The specific implementation process of the combined prediction includes: obtaining the 24-hour historical data of the new energy power generation unit; standardizing the 24-hour historical data to obtain normalized data; building an LSTM network, inputting the normalized data into the LSTM network to obtain the first prediction value; building an ARMA model, inputting the normalized data into the ARMA model to obtain the second prediction value; and combining the first prediction value and the second prediction value to obtain the new energy output prediction value by weighted average.

[0041] Step 3: Based on the communication network database of each node obtained in step 1, the database is divided into M equal time periods based on the maximum communication delay, and the selection probability of each time period is obtained based on the database;

[0042] Step 4: Based on the selection probability of each time period obtained in step 3, a power control instruction is issued based on the predicted value of the new energy output according to the roulette algorithm to compensate for the communication delay.

[0043] Furthermore, step one specifically includes the following steps:

[0044] Collect the delay data of each power node communication network and establish a database. The historical data recorded in the database is used to calculate the maximum communication delay of all node communication networks.

[0045] Furthermore, step 2 specifically includes the following steps:

[0046] Step 2-1: First, obtain the output historical data, meteorological historical data and date type data of the new energy power generation unit. The collected historical data is the data of one day in the distribution system.

[0047] Step 2-2: Normalize the historical data obtained in step 2-1 using the min-max normalization formula:

[0048]

[0049] where x ni represents the normalized data, x i represents the data to be normalized, x min Represents the minimum value of the data to be normalized, x max Indicates the maximum value of the data to be normalized.

[0050] Step 2-3: Design the LSTM network structure and establish an output prediction model for new energy power generation units based on LSTM. Specifically, LSTM uses input gates, forget gates, and output gates to protect and control information. Its specific functions are as follows:

[0051] f t =sigmoid(W f ·h t-1 +W f ·x t +b f ) (2)

[0052] i t =sigmoid(W i ·h t-1 +W i ·x t +b i ) (3)

[0053]

[0054]

[0055] o t =sigmoid(W o ·h t-1 +W o ·x t +b o ) (6)

[0056] h t =o t ⊙tanh(Ct ) (7)

[0057] where f t 、i t , o t Respectively represent the output signals of the forget gate, input gate, and output gate of the neuron; Represents the neuron candidate state information; C t Indicates neuron state information; h t Represents the hidden state of the neuron; sigmoid() is the S-type activation function; tanh() represents the hyperbolic tangent activation function; W f , W i , W C , W o Respectively represent the weights of the forget gate, input gate, memory unit and output gate; b f 、b i 、b C 、b o Respectively represent the bias of the forget gate, input gate, memory unit and output gate; x t represents the neuron input sequence value; ⊙ represents point-by-point multiplication operation.

[0058] Step 2-4: Train the LSTM network designed in step 2-3 to obtain the output prediction model of the renewable energy power generation unit. The output prediction model of the renewable energy power generation unit in step 2-4 is as follows: the training loss function is set to the root mean square error (RMSE), the number of iterations is set, and the Adam optimizer is selected to optimize the parameters in the network;

[0059]

[0060] Where n represents the number of prediction results; y i represents the predicted value; Represents the true value.

[0061] Step 2-5: The autoregressive model (AR) describes the relationship between the current value and the historical value, fits the data to establish a regression equation, and uses the historical time data of the variable itself to make predictions.

[0062] The p-order autoregressive process is shown in formula (9):

[0063]

[0064] Among them, y t and t-i They refer to the current value and the value at the previous i moments respectively; μ is a constant term, p refers to the order of the autoregressive model, and γ i refers to the autocorrelation coefficient, ε t It refers to the error at the current moment.

[0065] Step 2-6: The moving average model (MA) focuses on the accumulation of error terms in the autoregressive model, which can effectively eliminate random fluctuations in the forecast.

[0066] The q-order moving average process is shown in formula (10):

[0067]

[0068] Among them, ε t and ε t-i They refer to the error at the current moment and the error at the previous i moments respectively; μ is a constant term, q refers to the order of the moving average model, and θ i refers to the partial autocorrelation coefficient, y t Refers to the current value.

[0069] Step 2-7: Combine the autoregressive model and the moving average model by superposition to generate the following ARMA formula.

[0070] The combination of autoregression and moving average (ARMA) is shown in formula (11):

[0071]

[0072] Among them, y t and t-i They refer to the current value and the value at the previous i moments respectively; μ is a constant term, p refers to the order of the autoregressive model, and γ i refers to the autocorrelation coefficient, ε t and ε t-i They refer to the error at the current moment and the error at the previous i moments, respectively. q refers to the order of the moving average model. θ i Refers to the partial autocorrelation coefficient.

[0073] Step 2-8: Take the weighted average of the output forecast value of the new energy power generation unit obtained based on the LSTM network and the output forecast value of the new energy power generation unit obtained based on the ARMA algorithm to obtain the final output forecast value of the new energy power generation unit:

[0074]

[0075] Furthermore, step three specifically includes the following steps:

[0076] Step 3-1: Divide the maximum communication delay into M segments, namely:

[0077]

[0078] Step 3-2: Based on the communication delay database established in step 1, calculate the selection probability of each time period:

[0079]

[0080] Among them, P i The communication delay in the time period The probability of The communication delay in the time period The number of times, N τ is the total amount of communication delay data.

[0081] Furthermore, step four specifically includes the following steps:

[0082] Step 4-1: Calculate the selection probability of each communication delay time period:

[0083]

[0084] Among them, Q i is the selection probability of each communication delay time period.

[0085] Step 4-2: Select the compensation amount of communication delay based on the selection probability of each communication delay time period, and further select the communication delay compensation amount of each node based on the roulette algorithm And send the corresponding new energy output forecast value to each node according to the communication delay compensation amount:

[0086]

[0087] in, is the new energy output adjustment instruction of the jth node, is the first The predicted value at time.

[0088] This embodiment proposes a compensation method for distribution communication delay that takes into account the uncertainty of communication network delay and the randomness of renewable energy power generation. It improves the accuracy of renewable energy output power prediction by proposing a combined prediction algorithm based on LSTM and ARMA. It estimates the selection interval of communication network delay by calculating historical data and using a roulette algorithm. Finally, prediction compensation is used to solve the problem of the difficulty in accurately controlling renewable energy power caused by distribution network communication delay.

[0089] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A centralized compensation method for communication delay in a distribution network, characterized in that: include: Based on the historical delay data of the communication network of each power node in the distribution network, the maximum communication delay of all nodes is obtained; The LSTM network and ARMA model are used to combine and predict the power generation of the renewable energy power generation unit to obtain the predicted value of renewable energy output; Divide the maximum communication delay into several time periods, and obtain the selection probability of each time period based on the historical delay data; A power control instruction is formulated based on the selection probability of each time period and the predicted value of the new energy output.

2. The centralized compensation method for communication delay in distribution network according to claim 1, characterized in that: The process of using LSTM network and ARMA algorithm to combine and predict the power generation of new energy power generation units includes: Get 24-hour historical data of new energy power generation units; Standardizing the 24-hour historical data to obtain normalized data; Constructing an LSTM network, and inputting the normalized data into the LSTM network to obtain a first prediction value; Constructing an ARMA model, and inputting the normalized data into the ARMA model to obtain a second predicted value; The first prediction value and the second prediction value are combined and weighted averaged to obtain a new energy output prediction value.

3. The centralized compensation method for communication delay in distribution network according to claim 2, characterized in that: The 24-hour historical data of the new energy power generation unit obtained includes: output historical data, meteorological historical data and date type data.

4. The centralized compensation method for communication delay in distribution network according to claim 2, characterized in that: The process of building an ARMA model includes: combining the autoregressive model and the moving average model to obtain the ARMA model; wherein, the expression of the ARMA model is: In the formula, y t and t-i They refer to the current value and the value at the previous i moments respectively; μ is a constant term, p refers to the order of the autoregressive model, and γ i refers to the autocorrelation coefficient, ε t and ε t-i They refer to the error at the current moment and the error at the previous i moments, respectively. q refers to the order of the moving average model. θ i Refers to the partial autocorrelation coefficient.

5. The centralized compensation method for communication delay in distribution network according to claim 4, characterized in that: The expression for calculating the predicted value of new energy output is: In the formula, y i It refers to the predicted value of the LSTM network at the current moment, z t Represents the predicted value of new energy output.

6. The centralized compensation method for communication delay in distribution network according to claim 1, characterized in that: The expression of the selection probability in each time period is: Where P i The communication delay in the time period The probability of The communication delay in the time period The number of times, N τ is the total amount of communication delay data.

7. The centralized compensation method for communication delay in distribution network according to claim 1, characterized in that: The process of formulating a power control instruction and issuing a communication delay compensation amount based on the selection probability of each time period and the predicted output value of the new energy source includes: selecting the communication delay compensation amount of each node using a roulette algorithm based on the selection probability of each time period and the predicted output value of the new energy source; Among them, the expression of communication delay compensation is: In the formula, is the new energy output adjustment instruction of the jth node, is the first The predicted value at time.

Citation Information

Patent Citations

  • New energy power station primary frequency modulation networked control method

    CN113131532A