Data-driven TCPST emergency control start-up criterion generation method
Through a data-driven method, MATLAB and BPA interaction is used to obtain the fault limit clearing time, and a neural network is used to generate the TCPST emergency control startup criterion, which solves the problem of missing TCPST emergency control startup criterion, and improves the control response speed and service life. Ensures the stability and economic benefits of the power grid.
Patent Information
- Application Number
- CN202111125769.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-09-24
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2041-09-24
AI Technical Summary
In the prior art, there are few researches on transient control of TCPST, especially the lack of research on the initiation criteria for emergency control, which leads to frequent and unnecessary actions of TCPST during small disturbances or minor failures, affecting service life and increasing failure rate, and affecting the Steady-state control strategies and economic benefits of power systems.
Using a data-driven method, through the interaction of MATLAB and BPA, the fault limit clearing time under the fault location is automatically obtained, the sample set is learned using the neural network, and the emergency control start criterion of TCPST is generated, and the appropriate emergency control start criterion is selected. Improve the accuracy of machine learning based on the input volume of metrics and sample sets.
It improves the response speed and process framework integrity of TCPST emergency control, avoids the adverse impact of the lack of startup criteria on service life and grid benefits, and ensures effective control of TCPST in emergency situations.
Smart Images

Figure CN113852100B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of power systems, and in particular to a data-driven TCPST emergency control start criterion generation method and generation device. Background Art
[0002] Interconnected power systems have the advantages of high reliability of power generation and supply and strong operational flexibility. However, the transmission capacity of large systems is often limited by transient stability issues. Therefore, various control methods have been introduced into power systems to improve the transient stability of power systems and are applied to the control of various dynamic components of power systems. Since flexible AC transmission equipment including thyristor controlled phase shifters (TCPST) have the characteristics of fast response, they can not only be used for steady-state power flow control, but also can be used to suppress system oscillations and improve the transient stability of power systems.
[0003] At present, there are few studies on the transient control of TCPST, and there is no research on the TCPST emergency control start-up criteria. However, the start-up criteria are an indispensable part of the control framework of dynamic components. Without the start-up criteria, the emergency control of the component may be put into operation when there is a small disturbance or a minor fault, which will cause TCPST to act frequently and unnecessary, affecting the service life of TCPST related components, increasing the failure rate of the device, and may cause TCPST to be unable to maintain the optimal steady-state control strategy, thereby affecting TCPST from achieving the maximum economic benefit. Summary of the invention
[0004] The present application provides a data-driven TCPST emergency control start criterion generation method and generation device to solve the problems in the prior art that affect the service life of TCPST related components and increase the failure rate of devices.
[0005] In order to solve the above technical problems, the present application proposes a data-driven TCPST emergency control start criterion generation method, including: selecting the corresponding emergency control start criterion indicator according to the control target of TCPST; determining the input and output of the sample set used for data-driven; automatically obtaining the fault limit clearing time under different fault locations through the interaction between MATLAB and BPA; automatically obtaining sample sets with different fault locations and different fault clearing times through the interaction between MATLAB and BPA; learning the sample set through a neural network to obtain a learning model, and using the learning model as the emergency control start criterion of TCPST.
[0006] Optionally, according to the control target of TCPST, the corresponding emergency control start criterion index is selected, including: the OUT file output by the BPA transient stability simulation includes the maximum power angle difference in the system. When the maximum power angle difference is greater than 180 degrees, the system transient power angle is judged to be unstable, and this is used as the basis for judging the limit clearing time; where M = T CCT -T CT , T CCT is the limit clearing time, T CT is the fault clearing time, and M is the emergency control start criterion indicator.
[0007] Optionally, the input quantity of the sample set includes: generator power angle, node voltage and load node active power; the input quantity of the sample set is an emergency control start criterion indicator.
[0008] Alternatively, a dichotomy method is used to obtain the ultimate clearance time.
[0009] Optionally, the method further includes: importing the sample set into a neural network toolbox of MATLAB for learning, and improving the accuracy of machine learning by adjusting the number of samples and neural network parameters to obtain a learning model.
[0010] In order to solve the above technical problems, the present application proposes a data-driven TCPST emergency control start criterion generation device, including: an indicator module, used to select the corresponding emergency control start criterion indicator according to the control target of TCPST; a sample set module, used to determine the input and output of the sample set used for data-driven; a MATLAB and BPA module, used to automatically obtain the fault limit clearing time under different fault locations through the interaction between MATLAB and BPA; through the interaction between MATLAB and BPA, automatically obtain sample sets with different fault locations and different fault clearing times; a neural network module, used to learn the sample set through a neural network to obtain a learning model, and the learning model is used as the emergency control start criterion of TCPST.
[0011] Optionally, the OUT file output by the BPA transient stability simulation includes the maximum power angle difference in the system. When the maximum power angle difference is greater than 180 degrees, the system transient power angle is judged to be unstable, and this is used as the basis for judging the limit clearing time; wherein, M = T CCT -T CT , T CCT is the limit clearing time, T CT is the fault clearing time, and M is the emergency control start criterion indicator.
[0012] Optionally, the input quantity of the sample set includes: generator power angle, node voltage and load node active power; the input quantity of the sample set is an emergency control start criterion indicator.
[0013] Optionally, MATLAB and BPA modules are also used to obtain the limit clearance time using the bisection method.
[0014] Optionally, the network neural module is also used to import the sample set into the neural network toolbox of MATLAB for learning, and to improve the accuracy of machine learning by adjusting the number of samples and neural network parameters to obtain a learning model.
[0015] The present application proposes a data-driven TCPST emergency control start criterion generation method and generation device, the generation method includes selecting the corresponding emergency control start criterion indicator according to the control target of TCPST; determining the input and output of the sample set used for data driving; automatically obtaining the fault limit clearing time under different fault locations through the interaction between MATLAB and BPA; automatically obtaining sample sets with different fault locations and different fault clearing times through the interaction between MATLAB and BPA; learning the sample set through a neural network to obtain a learning model, and the learning model is used as the emergency control start criterion of TCPST. Through the above methods, the method of the present application has a fast calling speed, which helps to ensure the response speed of the control, and at the same time improves the process framework of TCPST emergency control, avoiding the adverse effects of the lack of start criteria on the service life of TCPST and the benefits of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the implementation manner will be briefly introduced below. Obviously, the drawings described below are only some implementation manners of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0017] Figure 1 It is a flow chart of an embodiment of a method for generating TCPST emergency control start criteria based on data-driven in the present application;
[0018] Figure 2 This is a system diagram of TCPST three machines and nine nodes in this embodiment;
[0019] Figure 3 This is a flowchart of an embodiment of step S130 of the present application;
[0020] Figure 4 This is a flowchart of an embodiment of step S140 of the present application;
[0021] Figure 5 It is a structural diagram of an embodiment of a data-driven TCPST emergency control start criterion generation device of the present application. DETAILED DESCRIPTION
[0022] In order to enable technical personnel in this field to better understand the technical solution of the present application, the data-driven TCPST emergency control start criterion generation method and generation device provided by the present application are further described in detail below in combination with the accompanying drawings and specific implementation methods.
[0023] In response to the lack of research on phase shifter emergency control starting criteria, this application proposes a data-driven TCPST emergency control starting criterion generation method, which improves the process framework of TCPST emergency control and avoids the adverse impact of the lack of starting criteria on the service life of TCPST and grid benefits.
[0024] In terms of software, the electromechanical transient simulation software package BPA developed by China Electric Power Research Institute has functions such as power flow calculation, transient stability simulation calculation, short circuit calculation, and small disturbance stability calculation. It is widely used in relevant units and universities such as the planning and design of my country's power system. Matrix Laboratory (MATLAB) is a commercial mathematical software launched by MathWorks for algorithm development, data visualization, data analysis, and numerical calculation. It is an advanced technical computing language and interactive environment. It has been widely used in many engineering fields. This patent is based on the interaction between BPA and MATLAB, combining the algorithm with the power system.
[0025] See also Figure 1 , Figure 1 This is a flow chart of an embodiment of a method for generating TCPST emergency control start criteria based on data-driven in this application. In this embodiment, the following steps may be specifically included:
[0026] S110: According to the control target of TCPST, select the corresponding emergency control start criterion indicator.
[0027] Step S110 is the selection and construction of indicators as the emergency control start-up criterion. Specifically, since the control target is the transient stability of the system, and considering that TCPST mainly controls the active power distribution of the branches, indicators related to the power angle stability should be selected. Considering that the OUT file output by the BPA transient stability simulation includes the maximum power angle difference in the system, when the maximum power angle difference is greater than 180 degrees, it can be judged that the system transient power angle is unstable, and this can be used as the basis for judging the limit clearing time;
[0028] Through the limit clearing time, the transient stability evaluation index is constructed as follows:
[0029] M=T CCT -T CT , where T CCT T is the critical clearing time (CCT),CT is the fault clearing time (CT), and M is the emergency control start criterion indicator.
[0030] See also Figure 2 , Figure 2 This is a system diagram of TCPST three machines and nine nodes in this embodiment. In the figure, 1-9 represent node positions, and G1-G3 represent generators.
[0031] S120: Determine the input amount and output amount of the sample set used for data driving.
[0032] Optionally, the input quantity of the sample set includes: generator power angle, node voltage and load node active power; the input quantity of the sample set is an emergency control start criterion indicator.
[0033] S130: Automatically obtain the fault limit clearing time at different fault locations through the interaction between MATLAB and BPA.
[0034] Optionally, use the bisection method to obtain the ultimate clearance time. Figure 3 , Figure 3 This is a flow chart of an embodiment of step S130 of the present application, specifically:
[0035] Fix a certain fault location. Set the initial lower limit of fault clearing time T a and the upper limit value T b , usually T a is 0, T b The value is selected according to the actual system, and generally a larger value is selected to ensure that T b There is a maximum power angle difference of >180 degrees.
[0036] If |T a -T b | is less than 0.0001s, then output T a The lower limit of the fault clearing time is set to T, and the next step is executed. Check whether all fault locations have been traversed. If yes, the simulation ends. If not, the fault location is changed and the initial lower limit of the fault clearing time is reset to T. a and the upper limit value T b .
[0037] If |T a -T b | Greater than or equal to 0.0001s, read the fault clearing time T CT =0.5*(T a +T b ), so as to judge whether the transient simulation is stable. Combined with the above, it can be seen that the result of transient stability can be judged by the maximum power angle difference. Specifically: if there is no maximum power angle difference>180 degrees, then the transient is stable, and 0.5*(Ta +T b ) is the new lower bound value T a And return to this step to judge |T a -T b | Is it less than 0.0001s? If the maximum power angle difference is >180 degrees, the transient state is unstable, and 0.5*(T a +T b ) is the new upper bound value T b And return to this step to judge |T a -T b |Is it less than 0.0001s?
[0038] The limit clearing times of different fault locations obtained in this embodiment are shown in Table 1.
[0039] Table 1 Limit clearing time of different fault locations
[0040]
[0041]
[0042] If all pre-set fault locations have been traversed, the simulation ends; otherwise, the fault location is changed and the previous step is returned.
[0043] S140: Through the interaction between MATLAB and BPA, sample sets with different fault locations and different fault clearing times are automatically obtained.
[0044] Specifically, see Figure 4 , Figure 4 1 is a flow chart of an embodiment of step S140 of the present application. Step S140 may include:
[0045] Fix a certain fault location and fault clearing time. Read the BPA transient model file (SWI file) through MATLAB code instructions, and write the fault location and fault clearing time into the SWI file. Execute the newly generated SWI file, obtain the result file (OUT file) through BPA transient simulation calculation, and read and save the OUT file data according to the needs of the sample set.
[0046] If all pre-set combinations of fault locations and fault clearing times have been traversed, the simulation ends; otherwise, the fault location and fault clearing time are changed, and the process returns to the above steps to re-write the fault location and fault clearing time into the SWI file.
[0047] S150: The sample set is learned through a neural network to obtain a learning model, and the learning model is used as an emergency control start criterion of TCPST.
[0048] Optionally, the method further includes: importing the sample set into the neural network toolbox of MATLAB for learning, and improving the accuracy of machine learning by adjusting the number of samples and neural network parameters to obtain a high-precision learning model.
[0049] The learning accuracy of the neural network is shown in Table 2, where MSE represents the mean square error of the model and R represents the correlation of the model. The sample set used for neural network learning is obtained at a load level of 100%, and the specific values corresponding to the load level of 100% are shown in Table 3. The subsequent verification is carried out at a load level of 105%. This is to illustrate that the startup criteria of this patent can adapt to the diversity of power system operation modes. The verification results are shown in Table 4. The specific verification scenario is: at a load level of 105%, the fault location is set to nodes 6 and 8, the fault clearing time is 0-20 cycles, the sampling period is 0.2 cycles, and the number of samples is 200.
[0050] Table 2 Neural network learning accuracy
[0051] Sample size MSE R Training samples 1574 2.90245e-4 9.99998e-1 Verification Sample 338 4.53455e-4 9.99997e-1 Test samples 338 3.58744e-4 9.99997e-1
[0052] Table 3 Load conditions of each load node when the system load level is 100%
[0053]
[0054]
[0055] Table 4 Verification results of the neural network model with a load level of 105%
[0056] Fault location Average error (cycle) Maximum clearing time (cycle) Error percentage (%) bus-6 0.1681 28.8154 0.58% bus-8 0.5745 18.6230 3.08%
[0057] When a fault occurs and is cleared after a certain period of time, the variable data corresponding to the input quantity of the sample set in step S120 at the moment of fault clearing is collected and input into the learning model of step S150. The learning model will quickly calculate the output quantity based on the input quantity. If the output quantity is greater than or equal to 0, the emergency control of TCPST will not be started. Otherwise, the emergency control of TCPST will be started.
[0058] As described above, the data-driven startup criteria proposed in this embodiment have a fast calling speed compared to traditional numerical calculation criteria, which helps to ensure the response speed of control; and it improves the process framework of TCPST emergency control, avoiding the adverse effects of the lack of startup criteria on the service life of TCPST and the benefits of the power grid.
[0059] Based on the above data-driven TCPST emergency control start criterion generation method, this application proposes a data-driven TCPST emergency control start criterion generation device, please refer to Figure 5 , Figure 5 It is a structural diagram of an embodiment of a data-driven TCPST emergency control start criterion generation device of the present application. In this embodiment, it may specifically include the following:
[0060] The indicator module is used to select the corresponding emergency control start criterion indicator according to the control target of TCPST;
[0061] A sample set module, used for determining the input amount and output amount of the sample set used for data driving;
[0062] MATLAB and BPA module, used to automatically obtain the fault limit clearing time under different fault locations through the interaction between MATLAB and BPA; automatically obtain sample sets of different fault locations and different fault clearing times through the interaction between MATLAB and BPA;
[0063] The neural network module is used to learn the sample set through the neural network to obtain a learning model, and the learning model is used as the emergency control start criterion of TCPST.
[0064] Optionally, the OUT file output by the BPA transient stability simulation includes the maximum power angle difference in the system. When the maximum power angle difference is greater than 180 degrees, the system transient power angle is judged to be unstable, and this is used as the basis for judging the limit clearing time; wherein, M = T CCT -T CT , T CCT is the limit clearing time, T CT is the fault clearing time, and M is the emergency control start criterion indicator.
[0065] Optionally, the input quantity of the sample set includes: generator power angle, node voltage and load node active power; the input quantity of the sample set is an emergency control start criterion indicator.
[0066] Optionally, MATLAB and BPA modules are also used to obtain the limit clearance time using the bisection method.
[0067] Optionally, the network neural module is also used to import the sample set into the neural network toolbox of MATLAB for learning, and to improve the accuracy of machine learning by adjusting the number of samples and neural network parameters to obtain a learning model.
[0068] The present application proposes a data-driven TCPST emergency control start criterion generation method and generation device, the generation method includes selecting the corresponding emergency control start criterion indicator according to the control target of TCPST; determining the input and output of the sample set used for data driving; automatically obtaining the fault limit clearing time under different fault locations through the interaction between MATLAB and BPA; automatically obtaining sample sets with different fault locations and different fault clearing times through the interaction between MATLAB and BPA; learning the sample set through a neural network to obtain a learning model, and the learning model is used as the emergency control start criterion of TCPST. Through the above methods, the method of the present application has a fast calling speed, which helps to ensure the response speed of the control, and at the same time improves the process framework of TCPST emergency control, avoiding the adverse effects of the lack of start criteria on the service life of TCPST and the benefits of the power grid.
[0069] It is to be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. In addition, for ease of description, only some structures related to the present application are shown in the accompanying drawings, rather than all structures. The step numbers used in the text are also only for the convenience of description and are not used as a limitation on the order in which the steps are executed. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.
[0070] The terms "first", "second", etc. in this application are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.
[0071] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0072] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly used in other related technical fields, are also included in the patent protection scope of the present application.
Claims
1. A data-driven TCPST emergency control start criterion generation method, characterized in that: include: According to the control objectives of TCPST, select the corresponding emergency control start criterion indicators; Determine the input and output quantities of the sample set used for data driving; Traversing a plurality of preset fault locations, and respectively setting the lower limit value of the fault clearing time of the currently traversed fault location to be 0 and the upper limit value to be a preset value, and then repeating the limit clearing time determination operation according to the lower limit value and the upper limit value until the fault limit clearing time of the currently traversed fault location is generated; Wherein, the fault clearing time of the currently traversed fault position is transiently unstable when the preset value is present; Through the interaction between MATLAB and BPA, sample sets with different fault locations and different fault clearing times are automatically obtained; The sample set is learned by a neural network to obtain a learning model, wherein the learning model is used as an emergency control start criterion of TCPST; The limit clearing time determination operation includes: Calculate the absolute value of the difference between the lower limit value and the upper limit value; wherein, initially, the lower limit value is 0 and the upper limit value is a preset value; When the absolute value is not less than 0.0001 seconds, the fault clearing time of the currently traversed fault position is set to half of the sum of the lower limit value and the upper limit value, and whether the transient simulation of the currently traversed fault position is stable is determined through BPA; If not, the current fault clearing time is determined as the upper limit value required for the next round of limit clearing time determination operation, and the next round of limit clearing time determination operation is performed; If yes, the current fault clearing time is determined as the lower limit required for the next round of limit clearing time determination operation, and the next round of limit clearing time determination operation is performed; When it is determined that the absolute value is less than 0.0001 seconds, the lower limit value is determined to be the fault limit clearing time of the current traversal fault position.
2. The data-driven TCPST emergency control start criterion generation method according to claim 1 is characterized in that: According to the control target of TCPST, the corresponding emergency control start criterion indicator is selected, including: The OUT file output by the BPA transient stability simulation includes the maximum power angle difference in the system. When the maximum power angle difference is greater than 180 degrees, the system transient power angle is judged to be unstable, and this is used as the basis for judging the limit clearing time. Where M = T CCT -T CT , T CCT is the limit clearing time, T CT is the fault clearing time, and M is the emergency control start criterion indicator.
3. The data-driven TCPST emergency control start criterion generation method according to claim 2 is characterized in that: The input quantity of the sample set includes: generator power angle, node voltage and load node active power; the output quantity of the sample set is the emergency control start criterion index.
4. The data-driven TCPST emergency control start criterion generation method according to claim 3 is characterized in that: Also includes: The sample set is imported into the neural network toolbox of the MATLAB for learning, and the accuracy of machine learning is improved by adjusting the number of samples and neural network parameters to obtain the learning model.
5. A data-driven TCPST emergency control start criterion generation device, characterized in that: include: The indicator module is used to select the corresponding emergency control start criterion indicator according to the control target of TCPST; A sample set module, used for determining the input amount and output amount of the sample set used for data driving; MATLAB and BPA modules, used to traverse a plurality of preset fault locations, and respectively set the lower limit value of the fault clearing time of the currently traversed fault location to 0 and the upper limit value to a preset value, and then repeat the limit clearing time determination operation according to the lower limit value and the upper limit value until the fault limit clearing time of the currently traversed fault location is generated; Wherein, the fault clearing time of the currently traversed fault position is transiently unstable when the preset value is present; Through the interaction between MATLAB and BPA, the sample sets of different fault locations and different fault clearing times are automatically obtained; A neural network module, used for learning the sample set through a neural network to obtain a learning model, wherein the learning model is used as an emergency control start criterion of TCPST; The limit clearing time determination operation includes: Calculate the absolute value of the difference between the lower limit value and the upper limit value; wherein, initially, the lower limit value is 0 and the upper limit value is a preset value; When the absolute value is not less than 0.0001 seconds, the fault clearing time of the currently traversed fault position is set to half of the sum of the lower limit value and the upper limit value, and whether the transient simulation of the currently traversed fault position is stable is determined through BPA; If not, the current fault clearing time is determined as the upper limit value required for the next round of limit clearing time determination operation, and the next round of limit clearing time determination operation is performed; If yes, the current fault clearing time is determined as the lower limit required for the next round of limit clearing time determination operation, and the next round of limit clearing time determination operation is performed; When it is determined that the absolute value is less than 0.0001 seconds, the lower limit value is determined to be the fault limit clearing time of the current traversal fault position.
6. The data-driven TCPST emergency control start criterion generation device according to claim 5 is characterized in that: The OUT file output by the BPA transient stability simulation includes the maximum power angle difference in the system. When the maximum power angle difference is greater than 180 degrees, the system transient power angle is judged to be unstable, and this is used as the basis for judging the limit clearing time. Where M = T CCT -T CT , T CCT is the limit clearing time, T CT is the fault clearing time, and M is the emergency control start criterion indicator.
7. The data-driven TCPST emergency control start criterion generation device according to claim 6 is characterized in that: The input quantity of the sample set includes: generator power angle, node voltage and load node active power; the input quantity of the sample set is the emergency control start criterion index.
8. The data-driven TCPST emergency control start criterion generation device according to claim 7 is characterized in that: The neural network module is also used to import the sample set into the neural network toolbox of the MATLAB for learning, and to improve the accuracy of machine learning by adjusting the number of samples and neural network parameters to obtain the learning model.
