A subsynchronous oscillation early warning strategy method based on finite state automata

Abstract

The present invention discloses a subsynchronous oscillation early warning strategy method based on a finite state automaton. According to the principle of subsynchronous control interaction, the method builds a Simulink simulation model of a doubly fed wind turbine, a series compensation capacitor, and a large power grid; then, according to the change trajectory of the three-phase total instantaneous power, a warning condition is set, a finite state automaton model is established, the power amplitude characteristics of the subsynchronous oscillation are identified, and an early warning element is written; then, a protection action element is set, and a suppression measure of bypassing the series compensation capacitor is taken when a subsynchronous oscillation early warning is issued; finally, the early warning constant is adjusted using Simulink and Matlab joint simulation until the set warning condition constant can effectively suppress the subsynchronous oscillation, and the adjustment result of the subsynchronous oscillation early warning condition constant is derived. The subsynchronous oscillation early warning strategy proposed by the present invention enables the power system to identify the occurrence of subsynchronous oscillation according to the change trajectory of the three-phase total instantaneous power, and the Simulink and Matlab joint simulation can accelerate the adjustment speed of the subsynchronous oscillation early warning condition constant.

Description

Technical Field

[0001] The present invention belongs to the field of subsynchronous oscillation monitoring and early warning of power systems, and in particular relates to a subsynchronous oscillation early warning strategy method based on a finite state automaton. Background Art

[0002] Subsynchronous oscillations (SSOs) are a type of oscillatory instability in power systems. They are caused by a unique electromechanical coupling in power systems. Their greatest threat is the potential for severe damage to the rotor shafting of large steam turbine generators, leading to major accidents and jeopardizing the safe operation of the power system. In recent years, the widespread application of power electronics has introduced interharmonics into the power grid, further increasing the likelihood of SSOs. The structure and grid connection methods of wind power generation systems differ fundamentally from those of traditional thermal power generators, and wind farms often consist of multiple types of wind turbines. This complicates the problem of SSOs in large-scale wind power generation.

[0003] Currently, there have been several studies on monitoring and controlling subsynchronous oscillations in large-scale renewable energy regions. For example, these studies use generator stator and rotor currents, or generator power variations, to identify subsynchronous oscillations. However, these algorithms for identifying subsynchronous oscillations are computationally complex and computationally intensive. Furthermore, the setting of warning thresholds often relies on multiple manual simulations, which is inefficient. Therefore, how to monitor and warn of subsynchronous oscillations and quickly set the warning thresholds has become an urgent problem. Summary of the Invention

[0004] The present invention aims to provide a subsynchronous oscillation early warning strategy based on a finite state automaton. This method can identify and warn of subsynchronous oscillations by comparing the trajectory of the three-phase total instantaneous power with the warning conditions. This method also accelerates the setting of the subsynchronous oscillation warning condition through co-simulation using Simulink and Matlab.

[0005] The object of the present invention is achieved through the following technical solutions:

[0006] A subsynchronous oscillation early warning strategy method based on a finite state automaton comprises the following steps:

[0007] (1) First, based on the subsynchronous oscillation identification technology and the change trajectory of the three-phase total instantaneous power, the corresponding instantaneous power waveform warning conditions are set;

[0008] (2) Identifying the power amplitude characteristics of subsynchronous oscillations and distinguishing the working conditions according to the warning conditions, and establishing a subsynchronous oscillation warning strategy based on a finite state automaton; the warning conditions of the instantaneous power waveform are the starting threshold conditions, power amplitude threshold conditions, power period threshold conditions, and oscillation number threshold conditions of the subsynchronous oscillation; the working conditions are synchronous oscillation and system power flow transfer, short circuit fault, or low-frequency oscillation;

[0009] (3) writing a subsynchronous oscillation warning element according to the warning condition of the instantaneous power waveform in step (1) and the subsynchronous oscillation warning strategy based on the finite state automaton described in step (2);

[0010] (4) Establish a power system simulation unit, a protection action element unit, and an early warning element unit, adjust the protection constants of the subsynchronous oscillation early warning strategy through Simulink simulation, and derive the adjustment results of the subsynchronous oscillation early warning condition constants.

[0011] Furthermore, the warning conditions of the instantaneous power waveform in step (1) are specifically:

[0012] According to the electrical quantity characteristic analysis of subsynchronous oscillation, both current and voltage contain subsynchronous components and supersynchronous components. The occurrence of subsynchronous oscillation can be determined by the comprehensive characteristics of power-representing current and voltage and the change trajectory of the three-phase total instantaneous power. The three-phase total instantaneous power P-factor calculation method is as follows:

[0013]

[0014] Among them, U m , I m are the amplitudes of voltage and current respectively, and θ is the phase angle of voltage leading current;

[0015] When the power grid is operating stably and normally, the three-phase current and three-phase voltage of each branch are symmetrically distributed in positive sequence, and the total instantaneous power P of the three phases is constant. After a subsynchronous oscillation occurs, P is no longer constant. The change trajectory of P can be used to determine the subsynchronous oscillation state. By setting four early warning conditions, the power amplitude characteristics of subsynchronous oscillation can be identified and distinguished from power system faults such as system power flow transfer, short circuit faults, and low-frequency oscillations.

[0016] The subsynchronous oscillation start threshold conditions are specifically:

[0017] |P-P0|>P qd

[0018] Among them, P is the instantaneous value of current power, P0 is the average power value within the sampling window time, and P qd It is the subsynchronous oscillation starting threshold, which is the warning value that needs to be adjusted;

[0019] The subsynchronous oscillation power amplitude threshold condition is specifically:

[0020] P max,k -P min,k >P dz

[0021] Among them, P max,k is the maximum power of the kth subsynchronous oscillation cycle, P min,k is the minimum power of the kth subsynchronous oscillation cycle, P dz is the power amplitude threshold of subsynchronous oscillation;

[0022] The subsynchronous oscillation power period threshold condition is specifically:

[0023] T max >T k >T min

[0024] Among them, T k is the period of the kth subsynchronous oscillation cycle, T max To identify the upper limit of the period threshold of subsynchronous oscillation, T min In order to identify the lower limit of the period threshold of subsynchronous oscillation, both are fixed values ​​that need to be adjusted;

[0025] The sub-synchronous oscillation number threshold condition is specifically:

[0026] N>N set

[0027] Where N is the number of oscillations of the current sub-synchronous oscillation, N set It is the oscillation number threshold of subsynchronous oscillation, which is a constant value that needs to be adjusted.

[0028] Furthermore, in step (2), the power amplitude characteristics of the subsynchronous oscillation are identified according to the warning conditions and the working conditions are distinguished, and a subsynchronous oscillation warning strategy based on a finite state automaton is established, specifically:

[0029] (3.1) Subsynchronous Oscillation Early Warning Strategy Based on Finite State Automata

[0030] The finite state automaton is a mathematical model of a finite number of states and the transitions and actions between these states, used to model object behavior and describe the sequence of states an object goes through during its life cycle, as well as how it responds to various events from the outside world.

[0031] (3.1.1) Define the state of the finite state automaton

[0032] To model the subsynchronous oscillation warning strategy as a finite state automaton, we first need to define eight different subsynchronous oscillation warning states:

[0033] State 0: The power system is in normal operation, and the active power waveform at the monitoring point does not trigger the subsynchronous oscillation starting threshold condition;

[0034] State 1: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold condition. However, because no complete oscillation cycle is subsequently monitored, the subsynchronous oscillation power amplitude, period, and oscillation number threshold conditions are not met.

[0035] State 2: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation number threshold condition is not met. The current subsynchronous oscillation number is 1.

[0036] State 3: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation number threshold condition is not met. The current subsynchronous oscillation number is 2.

[0037] State 4: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation count threshold condition is not met. The current subsynchronous oscillation count is 3.

[0038] State 5: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold condition, and the subsynchronous oscillation power amplitude, period, and oscillation number threshold conditions are all met, and an early warning is issued;

[0039] State 6: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold, but the subsynchronous oscillation power amplitude and period conditions are not met.

[0040] State 7: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition. The subsynchronous oscillation power amplitude and period conditions of the previous oscillation cycle have been met, but the subsynchronous oscillation power amplitude and period conditions of the most recent cycle have not been met.

[0041] (3.1.2) State transition of finite state automaton

[0042] The state transitions between eight different subsynchronous oscillation warning states are defined. When the monitoring device acquires a new sampling point, the state transition is triggered:

[0043] State 0 transfers to state 0: Currently in state 0, if the new sampling point data does not meet the sub-synchronous oscillation start threshold condition, it transfers to state 0;

[0044] State 0 transfers to state 1: Currently in state 0, if the new sampling point data meets the subsynchronous oscillation start threshold condition, it transfers to state 1;

[0045] State 1 transfer to state 1: Currently in state 1, if a complete oscillation cycle is not recorded when new sampling point data arrives, it will transfer to state 1;

[0046] State 1 transfers to State 2: Currently in State 1, if new sampling point data arrives, a complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then transfer to State 2;

[0047] State 1 transfers to State 6: Currently in State 1, if new sampling point data arrives and a complete oscillation cycle is recorded, and this cycle does not meet the subsynchronous oscillation amplitude threshold condition, or does not meet the subsynchronous oscillation frequency period threshold condition, then transfer to State 6;

[0048] State transfer to state 2: Currently in state 2, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 2;

[0049] State 2 to State 3: Currently in State 2, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 3;

[0050] State 2 transfers to State 7: Currently in State 2, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then transfer to State 7;

[0051] State 3 transfer to state 3: Currently in state 3, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 3;

[0052] State 3 to State 4: Currently in State 3, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 4;

[0053] State 3 transfers to State 7: Currently in State 3, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then the state transfers to State 7;

[0054] State 4 transfer to state 4: Currently in state 4, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 4;

[0055] State 4 to State 5: Currently in State 4, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 5.

[0056] State 4 transfers to State 7: Currently in State 4, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then transfer to State 7;

[0057] State 5 Transfer to State 5: Currently in State 5, if new sampling point data arrives and the warning is not released by manual intervention, it will transfer to State 5;

[0058] State 5 transfers to state 0: Currently in state 5, if new sampling point data arrives and the warning is released by manual intervention, it transfers to state 0;

[0059] State 6 transfer to state 6: Currently in state 6, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it transfers to state 6;

[0060] State 6 transfers to state 0: Currently in state 6, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data does not meet the subsynchronous oscillation amplitude threshold condition, or does not meet the subsynchronous oscillation period threshold condition, then transfer to state 0;

[0061] State 6 transfers to State 1: Currently in State 6, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data satisfies both the subsynchronous oscillation amplitude threshold condition and the period threshold condition, then the state transfers to State 1;

[0062] State 7 transfer to state 7: Currently in state 7, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 7;

[0063] State 7 transfers to state 0: Currently in state 7, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data does not meet the subsynchronous oscillation amplitude threshold condition, or does not meet the subsynchronous oscillation period threshold condition, then transfer to state 0;

[0064] State 7 transfers to State 1: Currently in State 7, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data satisfies both the subsynchronous oscillation amplitude threshold condition and the period threshold condition, then the state transfers to State 1;

[0065] (3.1.3) Entry actions of each state of the finite state machine

[0066] Then, the entry actions for 8 different subsynchronous oscillation warning states, that is, the actions to be performed when entering the state, are defined:

[0067] State 0: Clear the peak value record data, clear the peak value record flag, and set the oscillation number record to zero;

[0068] State 1: Clear the peak value record data, set the peak value record flag, and set the oscillation number record to zero;

[0069] State 2: The oscillation number record is assigned a value of 1;

[0070] State 3: The oscillation number record is assigned a value of 2;

[0071] State 4: The oscillation number record is assigned a value of 3;

[0072] State 5: The oscillation count is recorded as 4, and a subsynchronous oscillation warning is issued;

[0073] State 6: The oscillation number record is reset to zero;

[0074] State 7: The oscillation number record is reset to zero;

[0075] At this point, the complete definition of the finite state machine model of the subsynchronous oscillation early warning strategy is completed.

[0076] Furthermore, the step (3) of writing the subsynchronous oscillation warning element comprises the following specific steps:

[0077] (4.1) Subsynchronous oscillation warning component parameter settings

[0078] The subsynchronous oscillation warning element needs to collect the voltage waveform and current waveform of the doubly fed wind turbine and calculate the change trajectory of the three-phase total instantaneous power P. After obtaining the three-phase total instantaneous power P, the subsynchronous oscillation is warned according to the power amplitude characteristics of the subsynchronous oscillation and the subsynchronous oscillation warning strategy based on the finite state automaton. The sampling rate f of the warning element is set. s is 1KHz, the length of the sampling window is N wd The value is 1024. The parameters of the warning component include:

[0079] The starting threshold P of subsynchronous oscillation qd , is the fixed value of the warning condition (1) described in 2.1;

[0080] The power amplitude threshold P of subsynchronous oscillation dz , is the fixed value of the warning condition (2) described in 2.1;

[0081] The upper limit of the period threshold of subsynchronous oscillation T max , is the fixed value of the warning condition (3) described in 2.1;

[0082] The lower limit of the period threshold of subsynchronous oscillation T min , is the fixed value of the warning condition (3) described in 2.1;

[0083] The oscillation number threshold N of subsynchronous oscillation set , is the fixed value of the warning condition (4) described in 2.1;

[0084] (4.2) Simplified calculation of the average power P0 within the sampling window time

[0085] If each time the state transition occurs, assuming that the number of power sampling points in the current sampling window is N sample , and the average power of the sampling window in the previous state is P0′, which is divided into two cases:

[0086] If adding the current sampling point will make the number of sampling points exceed the sampling window size, that is, N sample +1>N wd , then you need to subtract the first sampling point A of the current sampling window t,P (1), plus the current sampling point P i , the equal signs in the following formulas all represent assignment operators, and the average power value P0 can be obtained by the following formula:

[0087]

[0088] If the number of sampling points does not exceed the sampling window size after adding the current sampling point, that is, N sample +1≤N wd , the average power value P0 can be obtained by the following formula, and N sample

[0089]

[0090] N sample =N sample +1

[0091] (4.3) Sampling window array A t,P Update

[0092] Each time the state transition occurs, a new sampling point P is obtained. i , then append it to A t,P At the end of the array, but since the size of the array is equal to the sampling window size Nwd , so if the number of data points in the array before appending is already equal to N wd , then you need to change the first element A of the array t,P (1) After deleting, add P i To A t,P The end of ; therefore, A t,P Essentially a size N wd Each data point in the sampling window contains two data, namely the time and power of the data point;

[0093] (4.4) Finite state automaton model for peak value judgment

[0094] The finite state automaton model is used in the power waveform peak judgment to prevent the power waveform peak from being misjudged;

[0095] (4.4.1) Finite automaton state S for peak value judgment rf_fsm definition

[0096] The state S of the state machine rf_fsm The value range is {-3,-2,-1,0,1,2,3}, and the specific definitions are as follows:

[0097] State 0: It is not determined whether it is currently on the rising edge or the falling edge;

[0098] State 1: Current sampling point P i The value is greater than the previous sampling point P i-1 The value of

[0099] State 2: Current sampling point power value P i Greater than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Greater than the power value P of the previous two sampling points i-2 ;

[0100] State 3: Current sampling point power value P i Greater than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Greater than the power value P of the previous two sampling points i-2 , and the power values ​​of the first two sampling points P i-2 Greater than the power value P of the first three sampling points i-3 , determine that the current power waveform is on the rising edge;

[0101] State-1: Current sampling point P i The value is less than the previous sampling point P i-1 The value of

[0102] State-2: Current sampling point power value Pi Less than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Less than the power value P of the previous two sampling points i-2 ;

[0103] State-3: Current sampling point power value P i Less than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Less than the power value P of the previous two sampling points i-2 , and the power values ​​of the first two sampling points P i-2 Less than the power value P of the first three sampling points i-3 , determine that the current power waveform is on the falling edge;

[0104] (4.4.2) Definition of Finite Automata State Transition for Peak Judgment

[0105] Then, the state transitions between the seven different peak judgment states are defined. When the monitoring device obtains a new sampling point, the state transition is triggered:

[0106] State 0 transfer to state 0: Currently in state 0, and the current sampling point is the first data point in the sampling window, transfer to state 0;

[0107] State 0 transfers to state 1: Currently in state 0, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 1;

[0108] State 1 transfers to state 2: Currently in state 1, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 2;

[0109] State 1 transfers to state 0: Currently in state 1, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0;

[0110] State 2 transfers to state 3: Currently in state 2, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 3;

[0111] State 2 transfers to state 0: Currently in state 2, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0;

[0112] State 3 transfers to state 3: Currently in state 3, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 3;

[0113] State 3 transfers to state 0: Currently in state 3, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0;

[0114] State 0 transfers to state -1: Currently in state 0, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -1;

[0115] State-1 transfers to state-2: Currently in state-1, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -2;

[0116] State -1 transfers to state 0: Currently in state -1, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0;

[0117] State-2 transfers to state-3: Currently in state-2, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -3;

[0118] State 2 transfers to state 0: Currently in state -2, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0;

[0119] State-3 transfer to state-3: currently in state-3, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -3;

[0120] State -3 transfers to state 0: Currently in state -3, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0;

[0121] (4.4.3) Finite state automaton entry action definition for peak value judgment

[0122] Then, the entry actions for different peak judgment states, that is, the actions performed when entering the state, are defined:

[0123] State-3: Set the rising edge / falling edge flag F rf =2, indicating that it is currently at the falling edge;

[0124] State 3: Set the rising edge / falling edge flag F rf =1, indicating that it is currently on the rising edge;

[0125] State {-2,-1,0,1,2}: Set the rising edge / falling edge flag F rf =F rf , indicating that the current flag remains unchanged;

[0126] So far, we have completed the complete definition of the finite state automaton model for peak judgment;

[0127] (4.5) Active power waveform peak value record array A peak Update

[0128] Array A peak The data addition is accompanied by the update of the rising / falling edge flag. Assume that the flag status before the update is F rf '=2, the updated flag status is F rf =1, it means that the minimum value is found, and the minimum value is the three sampling points before the current sampling point P i-3 , append it to array A peak ; Similarly, if the flag status before the update is F rf '=1, the updated flag status is F rf =2, it means that the maximum value is found, and the maximum value is the three sampling points before the current sampling point P i-3 , append it to array A peak middle;

[0129] Array A peak The data deletion is performed in the finite state machine of the sub-synchronization warning. fsm When transferring to state 0 or state 1, the corresponding entry action will be executed and array A will be cleared peak ; Complete the complete definition of subsynchronous oscillation warning components.

[0130] Specifically, in the step (4.1) of setting the subsynchronous oscillation warning component parameters, when the warning component executes the finite state machine model of the subsynchronous oscillation warning, the following data needs to be maintained:

[0131] The finite automaton state S of the current subsynchronous oscillation warning strategy fsm , its value range is {0,1,2,3,4,5,6,7};

[0132] Array A consisting of the sampling points in the sampling window t,P , the size of the array is N wd ;

[0133] The average power value P0 within the sampling window is equal to the sum of the power sampling points within the sampling window divided by the number of sampling points.

[0134] The oscillation count value N of the current synchronous oscillation;

[0135] Active power waveform peak value record array A peak ;

[0136] The index i of the peak record array currently traversed peak ;

[0137] Finite automaton state S for rising / falling edge judgment rf_fsm ;

[0138] The flag F of the current rising edge / falling edge rf .

[0139] Furthermore, in step (4), a power system simulation model is established, warning elements and protection action elements are set, and the protection setting values ​​of the subsynchronous oscillation warning strategy are adjusted through Simulink simulation, and the setting results of the subsynchronous oscillation warning condition setting values ​​are derived, specifically:

[0140] (5.1) Establishment of power system simulation model

[0141] Use the Simscape toolbox in the Simulink software component library, power system-specific component models in the electrical and professional power system directories, namely ideal voltage source, power line, transformer, RLC element, doubly fed wind turbine, synchronous generator, electric load, and circuit breaker models, and the Simulink oscilloscope model in the common component directory in the Simulink component library to build a power system simulation model that includes an infinite power grid, a doubly fed wind farm, and a series compensation line.

[0142] (5.2) Subsynchronous oscillation protection action element model

[0143] Based on subsynchronous oscillation suppression technology, a suppression measure is taken to bypass the series compensation capacitor when a subsynchronous oscillation warning is issued. Specifically, a circuit breaker connected in parallel at both ends of the series compensation capacitor simulates a protection action element to control the switching of the series compensation capacitor. When a subsynchronous oscillation warning is issued, the protection element is activated and closes the circuit breaker to bypass the series compensation capacitor.

[0144] (5.3) Subsynchronous Oscillation Warning Component Model

[0145] Based on the established subsynchronous oscillation warning element, the warning logic of the element is implemented through Matlab Function. When the element issues a warning, it will send a warning signal to the protection action element. The protection action will bypass the series compensation capacitor to suppress the subsynchronous oscillation;

[0146] (5.4) Setting the protection setting value of the subsynchronous oscillation warning strategy

[0147] The protection settings that need to be adjusted for the subsynchronous oscillation early warning strategy include:

[0148] The starting threshold P of subsynchronous oscillation qd , is the fixed value of the warning condition (1);

[0149] The power amplitude threshold P of subsynchronous oscillation dz , is the fixed value of the warning condition (2);

[0150] The upper limit of the period threshold of subsynchronous oscillation T max , is the fixed value of the warning condition (3);

[0151] The lower limit of the period threshold of subsynchronous oscillation T min , is the fixed value of the warning condition (3);

[0152] The oscillation number threshold N of subsynchronous oscillation set , is the fixed value of the warning condition (4);

[0153] Complete the parameter setting in the established subsynchronous oscillation warning component model, perform simulation, and analyze the simulation results; if the protection action under the warning condition can suppress the subsynchronous oscillation, then the parameter is valid; if the protection action under the warning condition cannot suppress the subsynchronous oscillation, then the warning condition needs to be tightened; for each set value, tightening the warning condition will increase the subsynchronous oscillation starting threshold P qd Increase, the power amplitude threshold P of subsynchronous oscillation dz Increase, the oscillation number threshold N of subsynchronous oscillation set Increase; on the contrary, if the warning conditions are relaxed, the starting threshold of subsynchronous oscillation P qd Reduce the power amplitude threshold P of subsynchronous oscillation dz Reduce the threshold N of the number of subsynchronous oscillations set Reduce; the upper limit of the period threshold T of the subsynchronous oscillation max and the lower limit of the period threshold of subsynchronous oscillation T min Determined by the frequency range of subsynchronous oscillation, no adjustment is required;

[0154] When adjusting the fixed value, the binary search method can shorten the adjustment process time; specifically, there are m fixed values ​​to be adjusted, and the i-th fixed value is vi (1≤i≤m), its upper limit is H i , the lower limit is L i , the adjustment amount is Δv i The upper limit, lower limit and adjustment amount are determined according to the physical meaning of the fixed value. During the adjustment process, the middle value m of the search interval of each fixed value is first calculated. i =(L i +H i ) / 2, then input the parameters into the warning element and perform simulation; analyze the simulation results. If the subsynchronous oscillation cannot be suppressed, the warning condition can be relaxed and H i =m i If the subsynchronous oscillation is suppressed, the warning condition can be tightened, so that L i =m i ; This process continues in a cycle until L i ≥H i , stop searching at this time and set the final setting value to v i =H i -Δv i , and then output the setting result; then the protection constant setting of the sub-synchronous oscillation early warning strategy is completed.

[0155] The beneficial effects of the present invention are as follows:

[0156] A subsynchronous oscillation early warning strategy based on the three-phase instantaneous power variation trajectory was implemented using a finite state automaton model. Compared to subsynchronous oscillation monitoring methods based on current and voltage, the power waveform-based early warning strategy offers the advantages of simpler algorithms and less computational effort. This finite state automaton-based early warning strategy accurately and sensitively determines the occurrence of subsynchronous oscillation. The proposed method for setting protection constants for the subsynchronous oscillation early warning strategy, based on Simulink simulation, enables rapid adjustment of the warning constants. BRIEF DESCRIPTION OF THE DRAWINGS

[0157] Figure 1 This is a state transition diagram of the sub-synchronous warning strategy based on the finite state automaton of the present invention;

[0158] Figure 2 This is a peak value judgment state transition diagram based on a finite state automaton of the present invention;

[0159] Figure 3 The figure is a schematic diagram of the protection constant value setting method of the subsynchronous oscillation early warning strategy based on Simulink simulation of the present invention. DETAILED DESCRIPTION

[0160] The present invention will be further described below with reference to the accompanying drawings.

[0161] A subsynchronous oscillation early warning strategy method based on a finite state automaton comprises the following steps:

[0162] (1) First, based on the subsynchronous oscillation identification technology and the change trajectory of the three-phase total instantaneous power, the corresponding instantaneous power waveform warning conditions are set;

[0163] (2) Identifying the power amplitude characteristics of subsynchronous oscillations and distinguishing the working conditions according to the warning conditions, and establishing a subsynchronous oscillation warning strategy based on a finite state automaton; the warning conditions of the instantaneous power waveform are the starting threshold conditions, power amplitude threshold conditions, power period threshold conditions, and oscillation number threshold conditions of the subsynchronous oscillation; the working conditions are synchronous oscillation and system power flow transfer, short circuit fault, or low-frequency oscillation;

[0164] (3) writing a subsynchronous oscillation warning element according to the warning condition of the instantaneous power waveform in step (1) and the subsynchronous oscillation warning strategy based on the finite state automaton described in step (2);

[0165] (4) Establish a power system simulation unit, a protection action element unit, and an early warning element unit, adjust the protection constants of the subsynchronous oscillation early warning strategy through Simulink simulation, and derive the adjustment results of the subsynchronous oscillation early warning condition constants.

[0166] like Figure 1-3 As shown, the present invention provides a subsynchronous oscillation early warning strategy based on a finite state automaton and a fixed value setting method thereof, comprising the following steps:

[0167] (1) Based on the subsynchronous oscillation identification technology, corresponding warning conditions are set according to the changing trajectory of the three-phase total instantaneous power.

[0168] (1.1) Power amplitude characteristics of subsynchronous oscillations

[0169] Analysis of the electrical quantity characteristics of subsynchronous oscillations shows that both current and voltage contain subsynchronous and supersynchronous components. Extracting subsynchronous oscillation characteristics from current alone is difficult to ensure identification sensitivity. Power, however, represents the combined characteristics of current and voltage. The change trajectory of the three-phase total instantaneous power can accurately and sensitively determine whether subsynchronous oscillations have occurred, and this method is easy to implement in embedded devices. The three-phase total instantaneous power P is calculated as follows:

[0170]

[0171] Among them, U m , I m are the amplitudes of voltage and current respectively, and θ is the phase angle of voltage leading current.

[0172] When the power grid is operating stably and normally, the three-phase currents and voltages of each branch are symmetrically distributed in positive sequence, and the total instantaneous power P of the three phases is constant. However, after a subsynchronous oscillation occurs, P is no longer constant. The change trajectory of P can be used to identify the subsynchronous oscillation state. The specific identification conditions are as follows:

[0173] (1.1.1) Subsynchronous Oscillation Start Threshold Conditions

[0174] |P-P0|>P qd

[0175] Among them, P is the current instantaneous power value, P0 is the average power value within the sampling window time, and P qd It is the subsynchronous oscillation starting threshold, which is an early warning value that needs to be adjusted.

[0176] (1.1.2) Subsynchronous Oscillation Power Amplitude Threshold Condition

[0177] P max,k -P min,k >P dz

[0178] Among them, P max,k is the maximum power of the kth subsynchronous oscillation cycle, P min,k is the minimum power of the kth subsynchronous oscillation cycle, P dz It is the power amplitude threshold of subsynchronous oscillation, that is, the constant value that needs to be adjusted.

[0179] (1.1.3) Subsynchronous Oscillation Power Cycle Threshold Condition

[0180] T max >T k >T min

[0181] Among them, T k is the period of the kth subsynchronous oscillation cycle, T max To identify the upper limit of the period threshold of subsynchronous oscillation, T min In order to identify the lower limit of the period threshold of subsynchronous oscillation, both are fixed values ​​that need to be adjusted.

[0182] (1.1.4) Subsynchronous Oscillation Number Threshold Condition

[0183] N>N set

[0184] Where N is the number of oscillations of the current sub-synchronous oscillation, N set It is the oscillation number threshold of subsynchronous oscillation, which is a constant value that needs to be adjusted.

[0185] By setting the above four conditions, the power amplitude characteristics of subsynchronous oscillations can be identified, effectively distinguishing subsynchronous oscillations from operating conditions such as system power flow transfer, short-circuit faults, and low-frequency oscillations, ensuring the reliability and sensitivity of monitoring and control performance.

[0186] (2) Establishing a subsynchronous oscillation early warning strategy based on finite state automata to accurately identify the power amplitude characteristics of subsynchronous oscillations, such as Figure 1 shown.

[0187] A finite state automaton is a mathematical model that represents a finite number of states and behaviors such as transitions and actions between these states. It is a tool used to model object behavior. Its main function is to describe the sequence of states that an object goes through during its life cycle and how it responds to various events from the outside world.

[0188] (2.1) State definition of finite state automaton

[0189] The subsynchronous oscillation warning strategy is modeled as a finite state automaton. First, eight different subsynchronous oscillation warning states are defined:

[0190] State 0: The power system is in normal operation, and the active power waveform at the monitoring point does not trigger the subsynchronous oscillation starting threshold condition;

[0191] State 1: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold condition. However, because no complete oscillation cycle is subsequently monitored, the subsynchronous oscillation power amplitude, period, and oscillation number threshold conditions are not met.

[0192] State 2: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation number threshold condition is not met. The current subsynchronous oscillation number is 1.

[0193] State 3: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation number threshold condition is not met. The current subsynchronous oscillation number is 2.

[0194] State 4: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation count threshold condition is not met. The current subsynchronous oscillation count is 3.

[0195] State 5: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold condition, and the subsynchronous oscillation power amplitude, period, and oscillation number threshold conditions are all met, and an early warning is issued;

[0196] State 6: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold, but the subsynchronous oscillation power amplitude and period conditions are not met.

[0197] State 7: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition. The subsynchronous oscillation power amplitude and period conditions of the previous oscillation cycle are met, but the subsynchronous oscillation power amplitude and period conditions of the most recent cycle are not met.

[0198] (2.2) State transition of finite state automaton

[0199] Then, the state transitions between the eight different subsynchronous oscillation warning states are defined. When the monitoring device obtains a new sampling point, the state transition is triggered. Specifically:

[0200] State 0 → State 0: Currently in state 0. If the new sampling point data does not meet the subsynchronous oscillation start threshold condition, it will transfer to state 0;

[0201] State 0 → State 1: Currently in State 0, if the new sampling point data meets the subsynchronous oscillation start threshold condition, it will transfer to State 1;

[0202] State 1 → State 1: Currently in State 1, if a complete oscillation cycle is not recorded when new sampling point data arrives, it will transfer to State 1;

[0203] State 1 → State 2: Currently in State 1, if new sampling point data arrives, a complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 2;

[0204] State 1 → State 6: Currently in State 1, if new sampling point data arrives and a complete oscillation cycle is recorded, and this cycle does not meet the subsynchronous oscillation amplitude threshold condition, or does not meet the subsynchronous oscillation frequency period threshold condition, then the state is transferred to State 6;

[0205] State 2 → State 2: Currently in State 2, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to State 2;

[0206] State 2 → State 3: Currently in State 2, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 3;

[0207] State 2 → State 7: Currently in State 2, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then the state will be transferred to State 7;

[0208] State 3 → State 3: Currently in State 3, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to State 3;

[0209] State 3 → State 4: Currently in State 3, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 4;

[0210] State 3 → State 7: Currently in State 3, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then the state transitions to State 7;

[0211] State 4 → State 4: Currently in State 4, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to State 4;

[0212] State 4 → State 5: Currently in State 4, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 5;

[0213] State 4 → State 7: Currently in State 4, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then the state will be transferred to State 7;

[0214] State 5 → State 5: Currently in State 5, if new sampling point data arrives and the warning is not released by manual intervention, it will transfer to State 5;

[0215] State 5 → State 0: Currently in State 5, if new sampling point data arrives and the warning is lifted by manual intervention, it will transfer to State 0;

[0216] State 6 → State 6: Currently in State 6. If a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to State 6.

[0217] State 6 → State 0: Currently in State 6, if new sampling point data arrives and a new complete oscillation cycle is recorded, and the cycle data does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation period threshold condition, then the state is transferred to State 0;

[0218] State 6 → State 1: Currently in State 6, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data meets both the subsynchronous oscillation amplitude threshold condition and the period threshold condition, then the state is transferred to State 1;

[0219] State 7 → State 7: Currently in state 7. If a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 7.

[0220] State 7 → State 0: Currently in State 7, if new sampling point data arrives and a new complete oscillation cycle is recorded, and the cycle data does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation period threshold condition, then it will transfer to State 0;

[0221] State 7 → State 1: Currently in State 7, if new sampling point data arrives and a new complete oscillation cycle is recorded, and the cycle data satisfies both the subsynchronous oscillation amplitude threshold condition and the period threshold condition, then the state is transferred to State 1.

[0222] (2.3) Entry actions of each state of the finite automatic state machine

[0223] Then, the entry actions for 8 different subsynchronous oscillation warning states, that is, the actions to be performed when entering the state, are defined:

[0224] State 0: Clear the peak value record data, clear the peak value record flag, and set the oscillation number record to zero;

[0225] State 1: Clear the peak value record data, set the peak value record flag, and set the oscillation number record to zero;

[0226] State 2: The oscillation number record is assigned a value of 1;

[0227] State 3: The oscillation number record is assigned a value of 2;

[0228] State 4: The oscillation number record is assigned a value of 3;

[0229] State 5: The oscillation count is recorded as 4, and a subsynchronous oscillation warning is issued;

[0230] State 6: The oscillation number record is reset to zero;

[0231] State 7: The oscillation number record is reset to zero.

[0232] (3) According to the warning conditions of the instantaneous power waveform and the subsynchronous oscillation warning strategy based on the finite state automaton, write the subsynchronous oscillation warning element, such as Figure 2 shown.

[0233] (3.1) Parameter definition of subsynchronous oscillation warning element

[0234] The subsynchronous oscillation warning component needs to collect the voltage and current waveforms of the doubly fed wind turbine and calculate the change trajectory of the three-phase total instantaneous power P. After obtaining the three-phase total instantaneous power P, the subsynchronous oscillation warning is issued according to the power amplitude characteristics of the subsynchronous oscillation and the subsynchronous oscillation warning strategy based on the finite state automaton. Set the sampling rate f of the warning component s is 1KHz, the length of the sampling window is N wd The value is 1024. The parameters of the warning component include:

[0235] The starting threshold P of subsynchronous oscillation qd , is the fixed value of the warning condition (1) described in 1.1;

[0236] The power amplitude threshold P of subsynchronous oscillation dz , is the fixed value of the warning condition (2) described in 1.1;

[0237] The upper limit of the period threshold of subsynchronous oscillation T max , is the fixed value of the warning condition (3) described in 1.1;

[0238] The lower limit of the period threshold of subsynchronous oscillation T min , is the fixed value of the warning condition (3) described in 1.1;

[0239] The oscillation number threshold N of subsynchronous oscillation set , is the fixed value of the warning condition (4) described in 1.1.

[0240] In addition, when the warning component executes the finite state machine model of subsynchronous oscillation warning, it needs to maintain the following data:

[0241] The finite automaton state S of the current subsynchronous oscillation warning strategy fsm , its value range is {0,1,2,3,4,5,6,7};

[0242] The (time, power) array A consisting of each sampling point in the sampling window t,P , the size of the array is N wd ;

[0243] The average power value P0 within the sampling window is equal to the sum of the power sampling points within the sampling window divided by the number of sampling points.

[0244] The oscillation count value N of the current synchronous oscillation;

[0245] Active power waveform peak value record array A peak ;

[0246] The index i of the peak record array currently traversed peak ;

[0247] Finite automaton state S for rising / falling edge judgment rf_fsm ;

[0248] The flag F of the current rising edge / falling edge rf .

[0249] (3.2) Simplified calculation of the average power P0 within the sampling window time

[0250] If the power sampling points in the sampling window are summed and then divided by the number of sampling points each time the state transition occurs, the amount of calculation is large. This process can be simplified. Assume that the number of power sampling points in the current sampling window is N sample , and the average power of the sampling window in the previous state is P0′, then we need to discuss two cases:

[0251] If adding the current sampling point will make the number of sampling points exceed the sampling window size, that is, N sample +1>N wd , then you need to subtract the first sampling point A of the current sampling window t,P (1), plus the current sampling point P i , the average power value P0 can be obtained by the following formula (unless otherwise specified, the equal sign in the formula below represents the assignment operator)

[0252]

[0253] If the number of sampling points does not exceed the sampling window size after adding the current sampling point, that is, N sample +1≤N wd , the average power value P0 can be obtained by the following formula, and N sample

[0254]

[0255] N sample =N sample +1

[0256] (3.3) Sampling window (time, power) array A t,P Update

[0257] Each time the state is transferred, a new sampling point P is obtained. i , then you need to append it to At,P At the end of the array, but since the size of the array is equal to the sampling window size N wd , so if the number of data points in the array before appending is already equal to N wd , then you need to change the first element A of the array t,P (1) After deleting, add P i To A t,P Therefore, A t,P Essentially a size N wd queue.

[0258] (3.4) Finite state automaton model for peak value judgment

[0259] In order to prevent misjudgment of the power waveform peak due to waveform jitter, glitches, etc., we apply the finite state automaton model to the power waveform peak judgment.

[0260] (3.4.1) Finite automaton state S for peak value judgment rf_fsm definition

[0261] The state S of the state machine rf_fsm The value range is {-3,-2,-1,0,1,2,3}, and the specific definitions are as follows:

[0262] State 0: It is not determined whether it is currently on the rising edge or the falling edge;

[0263] State 1: Current sampling point P i The value is greater than the previous sampling point P i-1 The value of

[0264] State 2: Current sampling point power value P i Greater than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Greater than the power value P of the previous two sampling points i-2 ;

[0265] State 3: Current sampling point power value P i Greater than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Greater than the power value P of the previous two sampling points i-2 , and the power values ​​of the first two sampling points P i-2 Greater than the power value P of the first three sampling points i-3 , determine that the current power waveform is on the rising edge;

[0266] State-1: Current sampling point P i The value is less than the previous sampling point P i-1 The value of

[0267] State-2: Current sampling point power value P i Less than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Less than the power value P of the previous two sampling points i-2 ;

[0268] State-3: Current sampling point power value P i Less than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Less than the power value P of the previous two sampling points i-2 , and the power values ​​of the first two sampling points P i-2 Less than the power value P of the first three sampling points i-3 , determine that the current power waveform is on the falling edge.

[0269] (3.4.2) Definition of Finite Automata State Transition for Peak Judgment

[0270] Then, the state transitions between the seven different peak judgment states are defined. When the monitoring device obtains a new sampling point, the state transition is triggered:

[0271] State 0 → State 0: Currently in state 0, and the current sampling point is the first data point in the sampling window, transfer to state 0;

[0272] State 0 → State 1: Currently in state 0, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 1;

[0273] State 1 → State 2: Currently in state 1, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 2;

[0274] State 1 → State 0: Currently in state 1, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0;

[0275] State 2 → State 3: Currently in state 2, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 3;

[0276] State 2 → State 0: Currently in state 2, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0;

[0277] State 3 → State 3: Currently in state 3, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 3;

[0278] State 3 → State 0: Currently in state 3, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0;

[0279] State 0 → State -1: Currently in state 0, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -1;

[0280] State -1→State -2: Currently in state -1, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -2;

[0281] State -1→State 0: Currently in state -1, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0;

[0282] State-2→State-3: Currently in state-2, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -3;

[0283] State 2 → State 0: Currently in state -2, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0;

[0284] State -3→State -3: Currently in state -3, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -3;

[0285] State -3 → State 0: Currently in state -3, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0;

[0286] (3.4.3) Finite state automaton entry action definition for peak value judgment

[0287] Then, the entry actions for different peak judgment states, that is, the actions performed when entering the state, are defined:

[0288] State-3: Set the rising edge / falling edge flag F rf =2, indicating that it is currently at the falling edge;

[0289] State 3: Set the rising edge / falling edge flag F rf =1, indicating that it is currently on the rising edge;

[0290] State {-2,-1,0,1,2}: Set the rising edge / falling edge flag F rf =F rf , indicating that the current flag remains unchanged;

[0291] So far, we have completed the complete definition of the finite state automaton model for peak judgment.

[0292] (3.5) Active power waveform peak value record array A peak Update

[0293] Array A peak The data addition is accompanied by the update of the rising / falling edge flag. Assume that the flag status before the update is F rf '=2, the updated flag status is F rf =1, it means that the minimum value is found, and the minimum value is the three sampling points before the current sampling point P i-3 , append it to array A peak ; Similarly, if the flag status before the update is F rf '=1, the updated flag status is F rf =2, it means that the maximum value is found, and the maximum value is the three sampling points before the current sampling point P i-3 , append it to array A peak middle.

[0294] Array A peak The data deletion is performed in the finite state machine of the sub-synchronization warning. fsm When transferring to state 0 or state 1, the corresponding entry action will be executed and array A will be cleared peak .

[0295] (4) Establish a power system simulation model, set the warning components and protection action components, adjust the protection setting value of the subsynchronous oscillation warning strategy through Simulink simulation, and derive the setting result of the subsynchronous oscillation warning condition setting value. The setting method of the warning condition setting value is as follows: Figure 3 shown.

[0296] (4.1) Building a power system simulation model

[0297] Using power system-specific component models such as ideal voltage sources, power lines, transformers, RLC elements, doubly-fed wind turbines, synchronous generators, electric loads, and circuit breaker models in the Simscape Toolbox / Electrical / Specialized Power Systems directory in the Simulink software component library, and oscilloscope models in the Simulink / Commonly Used Blocks directory in the Simulink component library, you can build power system simulation models that include infinite power grids, doubly-fed wind farms, series-compensated lines, etc.

[0298] (4.2) Constructing the subsynchronous oscillation protection action element model

[0299] Based on subsynchronous oscillation suppression technology, a suppression measure is implemented to bypass the series compensation capacitor when a subsynchronous oscillation warning is issued. Specifically, a circuit breaker connected in parallel across the series compensation capacitor simulates a protective action element to control the switching of the series compensation capacitor. When a subsynchronous oscillation warning is issued, the protective element is activated, closing the circuit breaker to bypass the series compensation capacitor.

[0300] (4.3) Constructing a subsynchronous oscillation warning component model

[0301] Based on the subsynchronous oscillation warning component model established in step 3, the warning logic of the component is implemented through Matlab Function. When the component issues a warning, it will send a warning signal to the protection action component. The protection action will bypass the series compensation capacitor to suppress subsynchronous oscillation.

[0302] (4.4) Setting the protection setting value of the subsynchronous oscillation warning strategy

[0303] The protection settings that need to be adjusted for the subsynchronous oscillation early warning strategy include: the subsynchronous oscillation starting threshold P qd , is the fixed value of the warning condition (1) described in 1.1; the power amplitude threshold P of the subsynchronous oscillation dz , is the fixed value of the warning condition (2) described in 1.1; the upper limit of the period threshold of subsynchronous oscillation T max , is the fixed value of the warning condition (3) described in 1.1; the lower limit of the period threshold of the subsynchronous oscillation T min , is the fixed value of the warning condition (3) described in 1.1; the oscillation number threshold N of subsynchronous oscillation set , is the fixed value of the warning condition (4) described in 2.1.

[0304] Complete the parameter setting in the subsynchronous oscillation warning element model established in step (4.3), perform simulation, and analyze the simulation results. If the protection action under the warning condition can suppress the subsynchronous oscillation, then the parameter is valid, but in order to meet the reliability requirements, that is, the protection only acts when it should act and does not act falsely when it should not act, it is necessary to relax the warning conditions; if the protection action under the warning condition cannot suppress the subsynchronous oscillation, in order to meet the speed requirements, it is necessary to tighten the warning conditions. For each set value, tightening the warning conditions means that the starting threshold P of the subsynchronous oscillation is less than 0. qd Increase, the power amplitude threshold P of subsynchronous oscillation dz Increase, the oscillation number threshold N of subsynchronous oscillation set Increase; on the contrary, relaxing the warning conditions means that the starting threshold of subsynchronous oscillation P qd Reduce the power amplitude threshold P of subsynchronous oscillation dz Reduce the threshold N of the number of subsynchronous oscillations set In addition, the upper limit of the period threshold of subsynchronous oscillation T max and the lower limit of the period threshold of subsynchronous oscillation T min It is determined by the frequency range of the subsynchronous oscillation and generally does not require adjustment.

[0305] When adjusting the fixed value, the binary search method can shorten the adjustment process time. Specifically, there are m fixed values ​​to be adjusted, and the i-th fixed value is v i (1≤i≤m), its upper limit is H i , the lower limit is L i , the adjustment amount is Δv i (The upper limit, lower limit and adjustment amount can be determined according to the physical meaning of the fixed value.) During the adjustment process, first calculate the middle value m of the search interval of each fixed value. i =(L i +H i ) / 2, then input the parameters into the warning component and perform simulation. Analyze the simulation results. If the subsynchronous oscillation cannot be suppressed, the warning condition can be relaxed and H i =m i If the subsynchronous oscillation is suppressed, the warning condition can be tightened, so that L i =m i This process continues in a cycle until L i ≥H i , stop searching at this time and set the final setting value to v i =H i -Δv i , and then output the tuning results.

[0306] For the subsynchronous oscillation starting threshold P qdand the power amplitude threshold P of subsynchronous oscillation dz , its lower limit Upper limit Adjustment value ΔP qd =0.1pu,ΔP dz =0.1pu; for the oscillation number threshold N of subsynchronous oscillation set , its lower limit Upper limit The adjustment value is ΔN set = 1. The adjustment process will continue until the critical value P of the protection setting is found. qd1 , P dz1 and N set1 Finally, the setting value of the subsynchronous oscillation starting threshold P qd_opt and the setting value of the power amplitude threshold P dz_opt , set it as the value minus a unit adjustment amount based on the critical value, that is,

[0307] P qd_opt =P qd1 -0.1

[0308] P dz_opt =P dz1 -0.1

[0309] N set_opt =N set -1

[0310] At this point, the protection setting value setting of the subsynchronous oscillation early warning strategy is completed.

[0311] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0312] The above embodiments are intended only to illustrate the design concepts and features of the present invention. Their purpose is to enable those skilled in the art to understand the contents of the present invention and implement them accordingly. The scope of protection of the present invention is not limited to the above embodiments. Therefore, any equivalent changes or modifications made based on the principles and design concepts disclosed in the present invention are within the scope of protection of the present invention.

Claims

1. A subsynchronous oscillation early warning strategy method based on finite state automata, characterized in that: The steps include: (1) First, based on the subsynchronous oscillation identification technology and the change trajectory of the three-phase total instantaneous power, the corresponding instantaneous power waveform warning conditions are set; the instantaneous power waveform warning conditions are specifically: According to the electrical quantity characteristic analysis of subsynchronous oscillation, both current and voltage contain subsynchronous components and supersynchronous components. The occurrence of subsynchronous oscillation can be determined by the comprehensive characteristics of power-representing current and voltage and the change trajectory of the three-phase total instantaneous power. The three-phase total instantaneous power P-factor calculation method is as follows: Among them, U m , I m are the amplitudes of voltage and current respectively, and θ is the phase angle of voltage leading current; When the power grid is operating stably and normally, the three-phase current and three-phase voltage of each branch are symmetrically distributed in positive sequence, and the total instantaneous power P of the three phases is constant. After a subsynchronous oscillation occurs, P is no longer constant. The change trajectory of P can be used to determine the subsynchronous oscillation state. By setting four early warning conditions, the power amplitude characteristics of subsynchronous oscillation can be identified and distinguished from power system faults such as system power flow transfer, short circuit faults, and low-frequency oscillations. The subsynchronous oscillation start threshold conditions are specifically: |P-P0|>P qd Among them, P is the instantaneous value of current power, P0 is the average power value within the sampling window time, and P qd It is the subsynchronous oscillation starting threshold, which is the warning value that needs to be adjusted; The subsynchronous oscillation power amplitude threshold condition is specifically: P max,k -P min,k >P dz Among them, P max,k is the maximum power of the kth subsynchronous oscillation cycle, P min,k is the minimum power of the kth subsynchronous oscillation cycle, P dz is the power amplitude threshold of subsynchronous oscillation; The subsynchronous oscillation power period threshold condition is specifically: T max >T k >T min Among them, T k is the period of the kth subsynchronous oscillation cycle, T max To identify the upper limit of the period threshold of subsynchronous oscillation, T min In order to identify the lower limit of the period threshold of subsynchronous oscillation, both are fixed values ​​that need to be adjusted; The sub-synchronous oscillation number threshold condition is specifically: N>N set Where N is the number of oscillations of the current sub-synchronous oscillation, N set The oscillation number threshold of subsynchronous oscillation is a fixed value that needs to be adjusted; (2) Identifying the power amplitude characteristics of subsynchronous oscillations and distinguishing the working conditions according to the warning conditions, and establishing a subsynchronous oscillation warning strategy based on a finite state automaton; the warning conditions of the instantaneous power waveform are the starting threshold conditions, power amplitude threshold conditions, power period threshold conditions and oscillation number threshold conditions of the subsynchronous oscillation; the working conditions are subsynchronous oscillation and system power flow transfer, short circuit fault or low frequency oscillation; (3) writing a subsynchronous oscillation warning element according to the warning condition of the instantaneous power waveform in step (1) and the subsynchronous oscillation warning strategy based on the finite state automaton described in step (2); (4) Establish a power system simulation unit, a protection action element unit, and an early warning element unit, adjust the protection constants of the subsynchronous oscillation early warning strategy through Simulink simulation, and derive the adjustment results of the subsynchronous oscillation early warning condition constants.

2. The subsynchronous oscillation early warning strategy method based on finite state automaton according to claim 1, characterized in that: The step (2) identifies the power amplitude characteristics of the subsynchronous oscillation according to the warning conditions and distinguishes the working conditions, and establishes a subsynchronous oscillation warning strategy based on a finite state automaton, specifically: (2.1) Subsynchronous Oscillation Early Warning Strategy Based on Finite State Automata The finite state automaton is a mathematical model of a finite number of states and the transitions and actions between these states, used to model object behavior and describe the sequence of states an object goes through during its life cycle, as well as how it responds to various events from the outside world. (2.1.1) Define the state of the finite state automaton To model the subsynchronous oscillation warning strategy as a finite state automaton, we first need to define eight different subsynchronous oscillation warning states: State 0: The power system is in normal operation, and the active power waveform at the monitoring point does not trigger the subsynchronous oscillation starting threshold condition; State 1: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold condition. However, because no complete oscillation cycle is subsequently monitored, the subsynchronous oscillation power amplitude, period, and oscillation number threshold conditions are not met. State 2: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation number threshold condition is not met. The current subsynchronous oscillation number is 1. State 3: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation number threshold condition is not met. The current subsynchronous oscillation number is 2. State 4: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition, and the subsynchronous oscillation power amplitude and period conditions are met, but the subsynchronous oscillation count threshold condition is not met. The current subsynchronous oscillation count is 3. State 5: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold condition, and the subsynchronous oscillation power amplitude, period, and oscillation number threshold conditions are all met, and an early warning is issued; State 6: The active power waveform at the monitoring point triggers the subsynchronous oscillation initiation threshold, but the subsynchronous oscillation power amplitude and period conditions are not met. State 7: The active power waveform at the monitoring point triggers the subsynchronous oscillation start threshold condition. The subsynchronous oscillation power amplitude and period conditions of the previous oscillation cycle have been met, but the subsynchronous oscillation power amplitude and period conditions of the most recent cycle have not been met. (2.1.2) State transition of finite state automaton The state transitions between eight different subsynchronous oscillation warning states are defined. When the monitoring device acquires a new sampling point, the state transition is triggered: State 0 transfers to state 0: Currently in state 0, if the new sampling point data does not meet the sub-synchronous oscillation start threshold condition, it transfers to state 0; State 0 transfers to state 1: Currently in state 0, if the new sampling point data meets the subsynchronous oscillation start threshold condition, it transfers to state 1; State 1 transfer to state 1: Currently in state 1, if a complete oscillation cycle is not recorded when new sampling point data arrives, it will transfer to state 1; State 1 transfers to State 2: Currently in State 1, if new sampling point data arrives, a complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then transfer to State 2; State 1 transfers to State 6: Currently in State 1, if new sampling point data arrives and a complete oscillation cycle is recorded, and this cycle does not meet the subsynchronous oscillation amplitude threshold condition, or does not meet the subsynchronous oscillation frequency period threshold condition, then transfer to State 6; State transfer to state 2: Currently in state 2, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 2; State 2 to State 3: Currently in State 2, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 3; State 2 transfers to State 7: Currently in State 2, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then transfer to State 7; State 3 transfer to state 3: Currently in state 3, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 3; State 3 to State 4: Currently in State 3, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 4; State 3 transfers to State 7: Currently in State 3, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then the state transfers to State 7; State 4 transfer to state 4: Currently in state 4, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 4; State 4 to State 5: Currently in State 4, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the maximum and minimum power values ​​within the cycle meet the subsynchronous oscillation amplitude threshold condition, and the oscillation period meets the subsynchronous oscillation period threshold condition, then the state is transferred to State 5. State 4 transfers to State 7: Currently in State 4, if new sampling point data arrives and a new complete oscillation cycle is recorded, but this cycle does not meet the subsynchronous oscillation amplitude threshold condition or the subsynchronous oscillation frequency period threshold condition, then transfer to State 7; State 5 Transfer to State 5: Currently in State 5, if new sampling point data arrives and the warning is not released by manual intervention, it will transfer to State 5; State 5 transfers to state 0: Currently in state 5, if new sampling point data arrives and the warning is released by manual intervention, it transfers to state 0; State 6 transfer to state 6: Currently in state 6, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it transfers to state 6; State 6 transfers to state 0: Currently in state 6, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data does not meet the subsynchronous oscillation amplitude threshold condition, or does not meet the subsynchronous oscillation period threshold condition, then transfer to state 0; State 6 transfers to State 1: Currently in State 6, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data satisfies both the subsynchronous oscillation amplitude threshold condition and the period threshold condition, then the state transfers to State 1; State 7 transfer to state 7: Currently in state 7, if a new sampling point data arrives and a new complete oscillation cycle is not recorded, it will transfer to state 7; State 7 transfers to state 0: Currently in state 7, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data does not meet the subsynchronous oscillation amplitude threshold condition, or does not meet the subsynchronous oscillation period threshold condition, then transfer to state 0; State 7 transfers to State 1: Currently in State 7, if new sampling point data arrives, a new complete oscillation cycle is recorded, and the cycle data satisfies both the subsynchronous oscillation amplitude threshold condition and the period threshold condition, then the state transfers to State 1; (2.1.3) Entry Actions of Each State of the Finite Automated State Machine Then, the entry actions for 8 different subsynchronous oscillation warning states, that is, the actions to be performed when entering the state, are defined: State 0: Clear the peak value record data, clear the peak value record flag, and set the oscillation number record to zero; State 1: Clear the peak value record data, set the peak value record flag, and set the oscillation number record to zero; State 2: The oscillation number record is assigned a value of 1; State 3: The oscillation number record is assigned a value of 2; State 4: The oscillation number record is assigned a value of 3; State 5: The oscillation count is recorded as 4, and a subsynchronous oscillation warning is issued; State 6: The oscillation number record is reset to zero; State 7: The oscillation number record is reset to zero; At this point, the complete definition of the finite state machine model of the subsynchronous oscillation early warning strategy is completed.

3. The subsynchronous oscillation early warning strategy method based on finite state automaton according to claim 1 is characterized in that: The specific steps of writing the subsynchronous oscillation warning component in step (3) are as follows: (3.1) Subsynchronous oscillation warning component parameter setting The subsynchronous oscillation warning component collects the voltage and current waveforms of the doubly fed wind turbine and calculates the change trajectory of the three-phase total instantaneous power P. After obtaining the three-phase total instantaneous power P, it issues a warning for subsynchronous oscillation based on the power amplitude characteristics of the subsynchronous oscillation and the subsynchronous oscillation warning strategy based on a finite state automaton. Set the sampling rate f of the warning element s is 1KHz, the length of the sampling window is N wd The value is 1024. The parameters of the warning component include: The starting threshold P of subsynchronous oscillation qd , is the fixed value of the warning condition (1) described in 2.1; The power amplitude threshold P of subsynchronous oscillation dz , is the fixed value of the warning condition (2) described in 2.1; The upper limit of the period threshold of subsynchronous oscillation T max , is the fixed value of the warning condition (3) described in 2.1; The lower limit of the period threshold of subsynchronous oscillation T min , is the fixed value of the warning condition (3) described in 2.1; The oscillation number threshold N of subsynchronous oscillation set , is the fixed value of the warning condition (4) described in 2.1; (3.2) Simplified calculation of the average power P0 within the sampling window time If each time the state transition occurs, assuming that the number of power sampling points in the current sampling window is N sample , and the average power of the sampling window in the previous state is P0′, which is divided into two cases: If adding the current sampling point will make the number of sampling points exceed the sampling window size, that is, N sample +1>N wd , then you need to subtract the first sampling point A of the current sampling window t,P (1), plus the current sampling point P i , the equal signs in the following formulas all represent assignment operators, and the average power value P0 can be obtained by the following formula: If the number of sampling points does not exceed the sampling window size after adding the current sampling point, that is, N sample +1≤N wd , the average power value P0 can be obtained by the following formula, and N sample N sample =N sample +1 (3.3) Sampling window array A t,P Update Each time the state transition occurs, a new sampling point P is obtained. i , then append it to A t,P At the end of the array, but since the size of the array is equal to the sampling window size N wd , so if the number of data points in the array before appending is already equal to N wd , then you need to change the first element A of the array t,P (1) After deleting, add P i To A t,P The end of ; therefore, A t,P Essentially a size N wd Each data point in the sampling window contains two data, namely the time and power of the data point; (3.4) Finite state automaton model for peak value judgment The finite state automaton model is used in the power waveform peak judgment to prevent the power waveform peak from being misjudged; (3.4.1) Finite automaton state S for peak value judgment rf_fsm definition The state S of the state machine rf_fsm The value range is {-3,-2,-1,0,1,2,3}, and the specific definitions are as follows: State 0: It is not determined whether it is currently on the rising edge or the falling edge; State 1: Current sampling point P i The value is greater than the previous sampling point P i-1 The value of State 2: Current sampling point power value P i Greater than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Greater than the power value P of the previous two sampling points i-2 ; State 3: Current sampling point power value P i Greater than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Greater than the power value P of the previous two sampling points i-2 , and the power values ​​of the first two sampling points P i-2 Greater than the power value P of the first three sampling points i-3 , determine that the current power waveform is on the rising edge; State-1: Current sampling point P i The value is less than the previous sampling point P i-1 The value of State-2: Current sampling point power value P i Less than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Less than the power value P of the previous two sampling points i-2 ; State-3: Current sampling point power value P i Less than the power value P of the previous sampling point i-1 , and the power value of the previous sampling point P i-1 Less than the power value P of the previous two sampling points i-2 , and the power values ​​of the first two sampling points P i-2 Less than the power value P of the first three sampling points i-3 , determine that the current power waveform is on the falling edge; (3.4.2) Definition of Finite Automata State Transition for Peak Judgment Then, the state transitions between the seven different peak judgment states are defined. When the monitoring device obtains a new sampling point, the state transition is triggered: State 0 transfer to state 0: Currently in state 0, and the current sampling point is the first data point in the sampling window, transfer to state 0; State 0 transfers to state 1: Currently in state 0, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 1; State 1 transfers to state 2: Currently in state 1, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 2; State 1 transfers to state 0: Currently in state 1, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0; State 2 transfers to state 3: Currently in state 2, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 3; State 2 transfers to state 0: Currently in state 2, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0; State 3 transfers to state 3: Currently in state 3, and the current sampling point P i The value is greater than the previous sampling point P i-1 , transfer to state 3; State 3 transfers to state 0: Currently in state 3, and the current sampling point P i The value is less than or equal to the previous sampling point P i-1 , transfer to state 0; State 0 transfers to state -1: Currently in state 0, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -1; State-1 transfers to state-2: Currently in state-1, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -2; State -1 transfers to state 0: Currently in state -1, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0; State-2 transfers to state-3: Currently in state-2, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -3; State 2 transfers to state 0: Currently in state -2, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0; State-3 transfer to state-3: currently in state-3, and the current sampling point P i The value is less than the previous sampling point P i-1 , transfer to state -3; State -3 transfers to state 0: Currently in state -3, and the current sampling point P i The value is greater than or equal to the previous sampling point P i-1 , transfer to state 0; (3.4.3) Finite state automaton entry action definition for peak value judgment Then, the entry actions for different peak judgment states, that is, the actions performed when entering the state, are defined: State-3: Set the rising edge / falling edge flag F rf =2, indicating that it is currently at the falling edge; State 3: Set the rising edge / falling edge flag F rf =1, indicating that it is currently on the rising edge; State {-2,-1,0,1,2}: Set the rising edge / falling edge flag F rf =F rf , indicating that the current flag remains unchanged; So far, we have completed the complete definition of the finite state automaton model for peak judgment; (3.5) Active power waveform peak value record array A peak Update Array A peak The data addition is accompanied by the update of the rising / falling edge flag. Assume that the flag status before the update is F rf '=2, the updated flag status is F rf =1, it means that the minimum value is found, and the minimum value is the three sampling points before the current sampling point P i-3 , append it to array A peak ; Similarly, if the flag status before the update is F rf '=1, the updated flag status is F rf =2, it means that the maximum value is found, and the maximum value is the three sampling points before the current sampling point P i-3 , append it to array A peak middle; Array A peak The data deletion is performed in the finite state machine of the sub-synchronization warning. fsm When transferring to state 0 or state 1, the corresponding entry action will be executed and array A will be cleared peak ; Complete the complete definition of subsynchronous oscillation warning components.

4. The subsynchronous oscillation early warning strategy method based on finite state automaton according to claim 3 is characterized in that: In the step (3.1) of setting the subsynchronous oscillation warning component parameters, when the warning component executes the finite state machine model of the subsynchronous oscillation warning, the following data needs to be maintained: The finite automaton state S of the current subsynchronous oscillation warning strategy fsm , its value range is {0,1,2,3,4,5,6,7}; Array A consisting of the sampling points in the sampling window t,P , the size of the array is N wd ; The average power value P0 within the sampling window is equal to the sum of the power sampling points within the sampling window divided by the number of sampling points. The oscillation count value N of the current synchronous oscillation; Active power waveform peak value record array A peak ; The index i of the peak record array currently traversed peak ; Finite automaton state S for rising / falling edge judgment rf_fsm ; The flag F of the current rising edge / falling edge rf .

5. The subsynchronous oscillation early warning strategy method based on finite state automaton according to claim 1, characterized in that: In step (4), a power system simulation model is established, warning elements and protection action elements are set, and the protection setting values ​​of the subsynchronous oscillation warning strategy are adjusted through Simulink simulation, and the setting results of the subsynchronous oscillation warning condition setting values ​​are derived, specifically: (4.1) Establishment of power system simulation model Use the Simscape toolbox in the Simulink software component library, power system-specific component models in the electrical and professional power system directories, namely ideal voltage source, power line, transformer, RLC element, doubly fed wind turbine, synchronous generator, electric load, and circuit breaker models, and the Simulink oscilloscope model in the common component directory in the Simulink component library to build a power system simulation model that includes an infinite power grid, a doubly fed wind farm, and a series compensation line. (4.2) Subsynchronous oscillation protection action element model Based on subsynchronous oscillation suppression technology, a suppression measure is taken to bypass the series compensation capacitor when a subsynchronous oscillation warning is issued. Specifically, a circuit breaker connected in parallel at both ends of the series compensation capacitor simulates a protection action element to control the switching of the series compensation capacitor. When a subsynchronous oscillation warning is issued, the protection element is activated and closes the circuit breaker to bypass the series compensation capacitor. (4.3) Subsynchronous Oscillation Warning Component Model Based on the established subsynchronous oscillation warning element, the warning logic of the element is implemented through Matlab Function. When the element issues a warning, it will send a warning signal to the protection action element. The protection action will bypass the series compensation capacitor to suppress the subsynchronous oscillation; (4.4) Setting the protection setting value of the subsynchronous oscillation warning strategy The protection settings that need to be adjusted for the subsynchronous oscillation early warning strategy include: The starting threshold P of subsynchronous oscillation qd , is the fixed value of the warning condition (1); The power amplitude threshold P of subsynchronous oscillation dz , is the fixed value of the warning condition (2); The upper limit of the period threshold of subsynchronous oscillation T max , is the fixed value of the warning condition (3); The lower limit of the period threshold of subsynchronous oscillation T min , is the fixed value of the warning condition (3); The oscillation number threshold N of subsynchronous oscillation set , is the fixed value of the warning condition (4); Complete the parameter setting in the established subsynchronous oscillation warning component model, perform simulation, and analyze the simulation results; if the protection action under the warning condition can suppress the subsynchronous oscillation, then the parameter is valid; if the protection action under the warning condition cannot suppress the subsynchronous oscillation, then the warning condition needs to be tightened; for each set value, tightening the warning condition will increase the subsynchronous oscillation starting threshold P qd Increase, the power amplitude threshold P of subsynchronous oscillation dz Increase, the oscillation number threshold N of subsynchronous oscillation set Increase; on the contrary, if the warning conditions are relaxed, the starting threshold of subsynchronous oscillation P qd Reduce the power amplitude threshold P of subsynchronous oscillation dz Reduce the threshold N of the number of subsynchronous oscillations set Reduce; the upper limit of the period threshold T of the subsynchronous oscillation max and the lower limit of the period threshold of subsynchronous oscillation T min Determined by the frequency range of subsynchronous oscillation, no adjustment is required; When adjusting the fixed value, the binary search method can shorten the adjustment process time; specifically, there are m fixed values ​​to be adjusted, and the i-th fixed value is v i (1≤i≤m), its upper limit is H i , the lower limit is L i , the adjustment amount is Δv i The upper limit, lower limit and adjustment amount are determined according to the physical meaning of the fixed value. During the adjustment process, the middle value m of the search interval of each fixed value is first calculated. i =(L i +H i ) / 2, then input the parameters into the warning element and perform simulation; analyze the simulation results. If the subsynchronous oscillation cannot be suppressed, the warning condition can be relaxed and H i =m i If the subsynchronous oscillation is suppressed, the warning condition can be tightened, so that L i =m i ; This process continues in a cycle until L i ≥H i , stop searching at this time and set the final setting value to v i =H i -Δv i , and then output the setting result; then the protection constant setting of the sub-synchronous oscillation early warning strategy is completed.

Images