Secondary frequency modulation intelligent control method based on auxiliary thermal power generating unit of energy storage system

Abstract

The invention discloses a secondary frequency modulation intelligent control method based on an auxiliary thermal power generating unit of an energy storage system, and the method comprises the following steps: decomposing an AGC instruction into an AGC main instruction and an AGC high-frequency instruction through employing a wavelet decomposition algorithm, and enabling the AGC main instruction to serve as a response instruction of the thermal power generating unit; inputting an AGC main instruction and an AGC high-frequency instruction to obtain an advanced multi-step prediction value, and constructing a prediction control system by using a particle filtering algorithm and a back-propagation neural network; an intelligent capacity control system is constructed, an original energy storage system is divided into a response energy storage system used for response, a self energy storage system used for energy storage and a supplementary energy storage system used for spare capacity adjustment, and the response energy storage system and the self energy storage system are fixed in capacity and form the intelligent capacity control system. The output response and the stabilizing response are respectively responsible for output response and stabilizing response. According to the technical scheme, the problem that the frequency modulation capability is poor due to capacity limitation when the fire-storage joint participates in the secondary frequency modulation of the power grid can be effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid secondary frequency regulation jointly participated by thermal energy storage, and in particular to an intelligent control method for secondary frequency regulation based on an energy storage system assisting a thermal power unit. Background Art

[0002] At present, clean energy such as wind power, photovoltaic power and hydropower accounts for nearly one-third of the total power generation of the national grid, and its intermittent power generation and insufficient cross-regional control capacity of the grid are becoming more prominent. Using energy storage systems to "eliminate peaks and flatten valleys" the nonlinear part of the AGC command after the clean energy is connected to the grid is the most mature and effective means of secondary frequency regulation in the current power grid, but the battery capacity directly limits its absorption performance. The energy storage system, namely the power system energy management system (Energy Management System, EMS), uses capacity intelligent control, which has become one of the important ways for energy storage systems to break through the battery capacity limit.

[0003] How to improve the maximum absorption capacity of the energy storage system under limited battery capacity is the key to building a capacity intelligent control system. Summary of the invention

[0004] In order to meet the above technical requirements, the purpose of the present invention is to provide an intelligent control method for secondary frequency regulation based on an energy storage system to assist thermal power units, which can effectively solve the problem of poor frequency regulation capability due to capacity limitations when thermal energy storage jointly participates in secondary frequency regulation of the power grid.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a secondary frequency regulation intelligent control method based on an energy storage system to assist a thermal power unit, comprising the following steps: Step 1) Decompose the AGC command into an AGC main command and an AGC high-frequency command using a wavelet decomposition algorithm, use the AGC main command as a thermal power unit response command, and use the AGC high-frequency command as an energy storage system response command; Step 2) Input the AGC main instruction and the AGC high-frequency instruction to obtain the multi-step-ahead prediction value, and use the particle filter algorithm and the back propagation neural network to build a predictive control system to complete the predictive control instruction output task; Step 3) Construct a capacity intelligent control system; divide the original energy storage system into a response energy storage system for response, a self-energy storage system for energy storage, and a spare energy storage system for spare capacity adjustment, wherein the response energy storage system and the self-energy storage system each have a fixed capacity, forming a capacity intelligent control system, which is responsible for output response and smooth response respectively; Step 4) The advanced multi-step AGC main command output by the prediction control system is brought into the thermal power unit for response, and the advanced multi-step AGC high-frequency command is responded by the capacity intelligent control system; Step 5) Perform polynomial fitting on the AGC response values ​​of the thermal power unit and the capacity intelligent control system and the original AGC command to obtain an error model, input the error into the spare energy storage system, and supplement the error to obtain a complete AGC compensation response; Step 6) At the same time, the AGC main response, AGC high frequency response and AGC compensation response are superimposed to complete the secondary frequency regulation intelligent control of the energy storage system assisting the thermal power unit.

[0006] Furthermore, the AGC main instruction in step 2) constructs a main instruction prediction system based on a particle filter algorithm, and the AGC high-frequency instructions are predicted by a nonlinear autoregressive neural network based on an external source input to obtain high-frequency AGC predicted instructions.

[0007] The capacity intelligent control system in step 3) is composed of multiple 18650 cylindrical batteries. Under the premise of a fixed total capacity, the capacity ratio of the response energy storage system and the own energy storage system can be allocated in advance according to the output instructions of the predictive control system.

[0008] In step 4), the responding energy storage system and the own energy storage system each function within their capacity carrying capacity. If the output or smoothing capacity exceeds the capacity carrying capacity of the responding energy storage system and the own energy storage system, capacity mutual assistance is performed.

[0009] The spare energy storage system in step 5) adjusts its capacity according to the multi-step-ahead prediction value to determine whether to respond or store energy. The capacity of the response or energy storage is the predicted result minus 150% of the capacity of its own energy storage system.

[0010] Through the above technical scheme, the present invention effectively improves the maximum absorption capacity of the energy storage system under limited battery capacity, successfully constructs a capacity intelligent control system, and ultimately solves the problem of poor frequency regulation capability due to capacity limitations when thermal energy storage jointly participates in secondary frequency regulation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a schematic diagram of the control method of the present invention.

[0012] Figure 2 It is a schematic diagram of capacity mutual assistance of the capacity intelligent control system in the present invention.

[0013] Figure 3 It is a schematic diagram of the secondary frequency modulation AGC response error in the present invention.

[0014] Figure 4 Schematic diagram of the AGC response error of the secondary frequency regulation of the traditional energy storage system assisted thermal power unit. DETAILED DESCRIPTION

[0015] See attached Figure 1The present invention fully demonstrates the secondary frequency regulation intelligent control method based on energy storage system to assist thermal power units. The implementation of this method uses 10,000 seconds of AGC command data of a thermal power plant in my country as the input data and frequency regulation performance evaluation standard of the thermal-storage combined secondary frequency regulation intelligent control method.

[0016] like Figure 2 The figure shows the capacity intelligent control system formed by the AGC high-frequency command response implemented by the present invention, which divides the fixed capacity into the response energy storage system, the energy storage system and the spare capacity. The AGC advance prediction command obtained by the AGC prediction control system is used to allocate the spare capacity, thereby improving the upper limit of the response energy storage system's absorption capacity.

[0017] Attached Figure 3 The AGC response error shown fully demonstrates the superior performance of the capacity intelligent control system.

[0018] Figure 4 Demonstrated the capacity intelligent control system and the traditional energy storage system to assist the secondary frequency regulation of thermal power units AGC response error comparison.

[0019] The thermal power unit model used in the present invention is a low-order model of five reheat steam turbines and a generator, and the number of participants is determined according to instruction requirements.

[0020] The above-mentioned secondary frequency regulation intelligent control method based on energy storage system assisting thermal power units is realized in the following ways: Step 1) Decompose the AGC command into an AGC main command and an AGC high-frequency command using a wavelet decomposition algorithm, use the AGC main command as a thermal power unit response command, and use the AGC high-frequency command as an energy storage system response command; Step 2) Input the AGC main instruction and the AGC high-frequency instruction to obtain the multi-step-ahead prediction value, and use the particle filter algorithm and the back propagation neural network to build a predictive control system to complete the predictive control instruction output task; Step 3) Construct a capacity intelligent control system; divide the original energy storage system into a response energy storage system for response, a self-energy storage system for energy storage, and a spare energy storage system for spare capacity adjustment, wherein the response energy storage system and the self-energy storage system each have a fixed capacity, forming a capacity intelligent control system, which is responsible for output response and smooth response respectively; Step 4) The advanced multi-step AGC main command output by the prediction control system is brought into the thermal power unit for response, and the advanced multi-step AGC high-frequency command is responded by the capacity intelligent control system; Step 5) Perform polynomial fitting on the AGC response values ​​of the thermal power unit and the capacity intelligent control system and the original AGC command to obtain an error model, input the error into the spare energy storage system, and supplement the error to obtain a complete AGC compensation response; Step 6) At the same time, the AGC main response, AGC high frequency response and AGC compensation response are superimposed to complete the secondary frequency regulation intelligent control of the energy storage system assisting the thermal power unit.

[0021] The implementation of the above method requires several models such as the predictive control system model, the capacity intelligent control system and the error model required for the spare energy storage system. The steps for building the predictive control system model are as follows:

[0022] Step 1) Decompose the 5000-second AGC command of a thermal power plant in my country into main command and high-frequency command by using COIF3 wavelet basis decomposition method, take the main prediction command as the response command of the thermal power unit, and take the high-frequency prediction command as the response command of the energy storage system.

[0023] Step 2) Use particle filter algorithm and back propagation neural network (BP neural network) to build a predictive control system. The basic steps are as follows: a) Extract the first 50% of the existing thermal power unit response instruction data as the training data set, and the remaining 50% as the verification data set; b) Using the training data set to train the standard particle filter algorithm model in MATLAB software, and using the verification data set to verify it, stop when the error between the single-step prediction result and the actual value is less than one percent, at which point the standard particle filter model under this parameter is considered to be the desired model; c) using a random sampling method to extract four data sets from the energy storage system response instructions at time intervals of 1 second, 3 seconds, 5 seconds and 7 seconds, namely, energy storage system response instruction data set 1, energy storage system response instruction data set 2, energy storage system response instruction data set 3 and energy storage system response instruction data set 4; d) Energy storage system response instruction data set 1, energy storage system response instruction data set 2, energy storage system response instruction data set 3 and energy storage system response instruction data set 4 are used to train the standard BP neural network in MATLAB software with the first two-thirds of the data, and the last one-third is used as the verification data to verify the model, requiring that the error between the predicted value and the actual value is not greater than one percent; e) Save the trained standard particle filter model and BP neural network.

[0024] Step 3) The standard particle filter model trained in step 2) is used as the AGC thermal power unit response prediction model, and the trained BP neural network model is used as the AGC energy storage system response model.

[0025] Step 1) The total capacity of the existing energy storage system is divided into a ratio of 2:2:6, and named as a response energy storage system, a self energy storage system and a spare energy storage system in sequence; Step 2) Input the 5000-second AGC high-frequency command data of a thermal power plant in my country into the AGC energy storage system response model (BP neural network model) in the predictive control system model to obtain the capacity control data set required by the capacity intelligent control system; Step 3) Use the first two-thirds of the capacity control data set as the training data set and the last one-third as the prediction data set to train and verify the standard particle filter algorithm model in MATLAB software. The verification accuracy error is required to be less than 5%. The parameters of the standard particle filter algorithm model are determined based on this accuracy condition. Step 4) Based on the prediction data output by the standard particle filter algorithm model in step 3, the charging and discharging operations are performed with the spare energy storage system as the basic energy storage capacity, and the proportion of the own energy storage system and the response energy storage system is dynamically adjusted according to the prediction data. Finally, the proportion of the power plant's own energy storage system, the response energy storage system and the spare energy storage system using the data is determined to be 1:3:6; Step 5) The capacity intelligent control system performs charging and discharging operations with the spare energy storage system as the basic capacity, supplements the energy storage capacity with its own energy storage system when the capacity exceeds the basic capacity of the spare energy storage system, and supplements the response capacity with the response energy storage system when the capacity exceeds the basic capacity of the spare energy storage system.

[0026] Step 1) The AGC main command response output by the thermal power unit and the AGC high-frequency command response output by the capacity intelligent control system are taken as the total AGC response of the thermal power plant, and the difference between the total AGC command response and the AGC command data of the thermal power plant is calculated, which is the AGC error compensation command; Step 2) The AGC error compensation instruction data is used as the error model input. The error model selects the double exponential fitting function, and the expression form of the function is selected as:

[0027] Where y is the predicted value, x is the AGC error compensation command input value, and a, b, c and d are the parameters required for fitting; Step 3) The 5000-second AGC command of a thermal power plant in my country is processed, and the obtained AGC main command and AGC high-frequency command are input into the predictive control system, and then passed through the thermal power unit and the capacity intelligent control system to obtain the uncompensated AGC total response, which is subtracted from the 5000-second AGC command of the thermal power plant to obtain the AGC error compensation command, and the AGC error compensation command data is used as the error model input to determine the double exponential fitting function a, b, c and d parameter values; Step 4) The double exponential fitting function for determining parameter values ​​is the error model required by the spare energy storage system of the present invention. Example

[0028] The 4500-second AGC instruction of a thermal power plant is used as the original input data to verify the performance of the thermal power plant combined with secondary frequency regulation intelligent control method proposed in the present invention. The performance verification simulation uses one-step-ahead prediction as the prediction step length, and all model output values ​​in the following steps are one-step-ahead prediction values.

[0029] Step 1: Decompose the 4500-second AGC command of a thermal power plant into an AGC main command and a high AGC frequency command by wavelet basis decomposition method, use the AGC main prediction command as the thermal power unit response command, and use the AGC high-frequency prediction command as the energy storage system response command; Step 2: The response command of the thermal power unit is input into the AGC thermal power unit response prediction model (standard particle filter model), and the response command of the energy storage system is input into the AGC energy storage system response model (BP neural network model), thereby obtaining the AGC response command of the thermal power unit and the AGC response high-frequency command of the capacity intelligent control system; Step 3: Input the AGC response command of the thermal power unit into the thermal power unit to obtain the AGC response of the thermal power unit, input the AGC response high-frequency command of the capacity intelligent control system into the capacity intelligent control system to obtain the AGC response of the capacity intelligent control system, input the sum of the two response values ​​into the error model (double exponential fitting function), and output the error model to the spare energy storage system to obtain the AGC supplementary response.

[0030] Step 4: AGC response of thermal power unit, the sum of AGC response of capacity intelligent control system and AGC supplementary response is the one-step-ahead response value of the thermal power unit combined secondary frequency regulation intelligent control method proposed by the present invention. Attached Figure 3 This is a schematic diagram of the AGC response error of this example.

Claims

1. A secondary frequency regulation intelligent control method based on energy storage system to assist thermal power generation units, characterized in that The following steps are involved: Step 1) Decomposing the AGC instruction into an AGC main instruction and an AGC high-frequency instruction by using a wavelet decomposition algorithm, taking the AGC main instruction as a thermal power unit response instruction, and taking the AGC high-frequency instruction as an energy storage system response instruction; Step 2) Input the AGC main instruction and the AGC high-frequency instruction to obtain the multi-step-ahead prediction value, and use the particle filter algorithm and the back propagation neural network to build a predictive control system to complete the predictive control instruction output task; Step 3) constructing a capacity intelligent control system; dividing the original energy storage system into a response energy storage system for response, a self-energy storage system for energy storage, and a supplementary energy storage system for spare capacity adjustment, wherein the response energy storage system and the self-energy storage system each have a fixed capacity, forming a capacity intelligent control system, which are responsible for output response and smooth response respectively; Step 4) The AGC main command output by the prediction control system is brought into the thermal power unit for response, and the AGC high-frequency command is responded by the capacity intelligent control system; Step 5) Perform polynomial fitting on the AGC response value and the original AGC command to obtain an error model, and the supplementary energy storage system supplements the error to obtain a complete AGC response.

2. The secondary frequency regulation intelligent control method based on energy storage system assisting thermal power generation units according to claim 1 is characterized by: The AGC main instruction in step 2) constructs a main instruction prediction system based on a particle filter algorithm, and the AGC high-frequency instructions are predicted by a nonlinear autoregressive neural network based on external input to obtain high-frequency AGC predicted instructions.

3. The secondary frequency regulation intelligent control method based on energy storage system assisting thermal power generation units according to claim 1 is characterized by: The capacity intelligent control system in step 3) is composed of multiple 18650 cylindrical batteries. Under the premise of a fixed total capacity, the capacity ratio of the response energy storage system and the own energy storage system can be allocated in advance according to the output instructions of the predictive control system.

4. The secondary frequency regulation intelligent control method based on energy storage system assisting thermal power generation units according to claim 1 is characterized by: In step 4), the responding energy storage system and the own energy storage system each function within their capacity carrying capacity. If the output or smoothing capacity exceeds the capacity carrying capacity of the responding energy storage system and the own energy storage system, capacity mutual assistance is performed.

5. The secondary frequency regulation intelligent control method based on energy storage system assisting thermal power generation units according to claim 1 is characterized by: The supplementary energy storage system in step 5) adjusts its capacity according to the multi-step-ahead prediction value to determine whether to respond or store energy. The capacity of the response or energy storage is the predicted result minus 150% of the capacity of its own energy storage system.

Images