CN108365615B - Self-adaptive wide area damping controller and control method - Google Patents
Self-adaptive wide area damping controller and control method Download PDFInfo
- Publication number
- CN108365615B CN108365615B CN201810126049.XA CN201810126049A CN108365615B CN 108365615 B CN108365615 B CN 108365615B CN 201810126049 A CN201810126049 A CN 201810126049A CN 108365615 B CN108365615 B CN 108365615B
- Authority
- CN
- China
- Prior art keywords
- signal
- adaptive
- time lag
- phase
- self
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000013016 damping Methods 0.000 title claims abstract description 35
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000004891 communication Methods 0.000 claims abstract description 53
- 230000010355 oscillation Effects 0.000 claims abstract description 52
- 230000003044 adaptive effect Effects 0.000 claims abstract description 49
- 238000005259 measurement Methods 0.000 claims abstract description 33
- 238000013528 artificial neural network Methods 0.000 claims abstract description 30
- 230000010363 phase shift Effects 0.000 claims abstract description 18
- 230000006870 function Effects 0.000 claims description 46
- 238000012546 transfer Methods 0.000 claims description 33
- 230000000694 effects Effects 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 11
- 230000006399 behavior Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 230000002401 inhibitory effect Effects 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims 1
- 230000001052 transient effect Effects 0.000 abstract description 16
- 230000003321 amplification Effects 0.000 abstract 1
- 230000001934 delay Effects 0.000 abstract 1
- 238000003199 nucleic acid amplification method Methods 0.000 abstract 1
- 230000005540 biological transmission Effects 0.000 description 10
- 230000008859 change Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 10
- 230000002787 reinforcement Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 238000012937 correction Methods 0.000 description 3
- 238000013178 mathematical model Methods 0.000 description 3
- 230000001629 suppression Effects 0.000 description 3
- 230000006872 improvement Effects 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/24—Arrangements for preventing or reducing oscillations of power in networks
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a self-adaptive wide area damping controller and a control method, wherein the controller comprises: the device comprises a self-adaptive time lag compensator, a phase shift unit and a GrHDP unit; the self-adaptive time lag compensator is used for carrying out self-adaptive time lag compensation on the wide area measurement signal to obtain a signal x (t); the phase shifting unit is used for amplifying and shifting the phase of the signal X (t) to obtain a parallel phase-shifted signal X (t); the GrHDP unit obtains a control signal u (t) adaptive to the current operation condition of the power system according to the parallel phase-shifting signal X (t) based on self-adaptive dynamic programming; the control method comprises the following steps: (1) performing adaptive time lag compensation on the wide area measurement signal; (2) obtaining parallel phase-shifted signals through amplification and phase shifting; (3) and obtaining a control signal u (t) adaptive to the current operation condition of the power system by utilizing the GrHDP neural network. The invention can effectively inhibit the low-frequency oscillation of the system under different operation conditions and different communication time delays, and improves the transient stability of the system.
Description
Technical Field
The invention belongs to the field of flexible direct current power systems, and particularly relates to a self-adaptive wide area damping controller and a control method.
Background
The flexible direct current transmission technology (VSC-HVDC) has the advantages of flexible control, no phase commutation failure risk, independent reactive power control and the like, and is one of key technologies for new energy grid connection and power supply to a passive system. By adopting a back-to-back flexible direct current transmission technology (BTB-VSC-HVDC), the interconnection of two asynchronously-operated alternating current power grids can be realized, so that the operation controllability of the power grids is improved, and the safe and stable operation risk of the power grids is reduced. By combining with a wide-area measurement system (WAMS) technology, a wide-area damping controller (WADC) is reasonably designed for the back-to-back flexible direct-current power system, and low-frequency oscillation of the power system can be effectively inhibited.
At present, a conventional VSC-HVDC (voltage source-area damping controller, C-WADC) is designed by utilizing a linear mathematical model of a system under a certain typical operation condition, the adaptability to the operation condition of system change is poor, and an accurate mathematical model of an actual power system is difficult to obtain.
In addition, communication time lag inevitably exists in the transmission process of the wide area measurement signal, the time lag influence is not considered in the WADC design in the past, or the time lag is considered to be fixed, however, when the system is disturbed differently, the time lag can be changed; the time-varying skew may cause the performance of the WADC control to degrade and may even threaten the transient stability of the system.
Therefore, it is necessary to adopt a model-independent damping controller to perform adaptive control for different operating conditions, suppress low-frequency oscillation of the system, and provide adaptive compensation capability for communication time lag, so as to perform corresponding compensation for different communication time lags.
Disclosure of Invention
Aiming at the defects and improvement requirements of the prior art, the invention provides a self-adaptive wide area damping controller and a control method, and aims to effectively inhibit low-frequency oscillation of a system and improve the transient stability of the system under different operating conditions and different communication time lags.
To achieve the above object, according to a first aspect of the present invention, there is provided an adaptive wide area damping controller comprising: the device comprises a self-adaptive time lag compensator, a phase shift unit and a GrHDP unit;
an adaptive-delay compensator (ADC) input for receiving wide-area measurement signals
Adaptive skew compensator for wide area measurement signals to an electrical power system
Carrying out self-adaptive time lag compensation to obtain a signal x (t);
the input end of the phase shifting unit is connected to the output end of the self-adaptive time lag compensator, and the phase shifting unit is used for amplifying and phase-shifting the signal X (t) to obtain a parallel phase-shifted signal X (t);
the input end of the GrHDP unit is connected to the output end of the phase shifting unit, and the GrHDP unit is used for obtaining a control signal u (t) adaptive to the current operation condition of the power system according to the parallel phase shifting signal X (t) so as to realize the adaptive compensation of active power oscillation and reactive power oscillation of the power system, and further effectively inhibit the low-frequency oscillation of the power system.
Further, the adaptive skew compensator includes n skew compensation submodules (sub-delay compensators, SDC); the transfer function of the adaptive skew compensator is a weighted sum of the transfer functions of n skew compensation submodules, and the weight of each skew compensation submodule is combined with the wide-area measurement signal
The communication time lag tau is related, so that the self-adaptive time lag compensator can correspondingly compensate different communication time lags, and the communication time lag measurement signal in a wide area can be effectively eliminated
The lag phase introduced therein; the number n of the time lag compensation submodules is determined according to the maximum communication time lag of the system, and a trial and error method is generally adopted, and 4-8 time lags are selected.
Further, the phase shift unit comprises a first amplifier, a second amplifier and a phase shifter; the first amplifier is used for amplifying the signal x (t) by k1Multiplying to obtain a first path of signal; a second amplifier for amplifying the signal x (t) by k2Multiplying to obtain an intermediate signal x' (t); the input end of the phase shifter is connected to the output end of the second amplifier, and the phase shifter is used for shifting the phase of the intermediate signal x' (t) to obtain a second path of signal; the parallel phase shift signal X (t) comprises a first path of signal and a second path of signal;
further, the mathematical expression of the phase shifter is
Wherein, TfThe filter constant is used for preventing the differential link from amplifying high-frequency noise to influence the control effect, and the value range of the filter constant is 0.01-0.05;
further, k is1And k2The normalization coefficient is used for ensuring that the amplitude ranges of the parallel first path of signals and the second path of signals are consistent.
Further, the GrHDP unit realizes the calculation from the parallel phase-shifting signal X (t) to the control signal u (t) based on the self-adaptive dynamic programming algorithm; the control signal u (t) includes an additional active power command value Δ Pref(t) and an additional reactive power command value DeltaQref(t); additional active power command value Δ Pref(t) for compensating for oscillations in active power, adding a reactive power command value DeltaQref(t) for compensating oscillations of the reactive power.
According to a second aspect of the present invention, there is provided a control method based on the adaptive wide-area damping controller provided by the first aspect of the present invention, comprising the steps of:
(1) for wide area measurement signal
Carrying out self-adaptive time lag compensation to obtain a signal x (t);
(2) amplifying and phase-shifting the signal X (t) to obtain a parallel phase-shifted signal X (t);
(3) and obtaining a control signal u (t) adaptive to the current operating environment of the power grid according to the parallel phase-shifting signal X (t) so as to realize the self-adaptive compensation of the active power oscillation and the reactive power oscillation of the power grid, thereby effectively inhibiting the low-frequency oscillation of the power system.
Further, the step (1) comprises the following steps:
(11) computing wide area measurement signals
Time-lag transfer function G ofd(s); and transfer function G is transformed by using second-order Pade approximationd(s) to obtain a transfer function GD(s), transfer function GDThe calculation formula of(s) is as follows:
wherein tau is a wide-area measurement signal
Communication skew of (2);
(12) calculating a transfer function and a corresponding weight of each time-lag compensation submodule; the transfer function calculation formula of each time lag compensation submodule is as follows:
wherein, TcIs a time constant associated with the steady state behavior of the system; taking into account the steady-state behavior of the system, TcHas a value range of [0.01s,0.1s ]]；
The weight calculation formula of the time lag compensation submodule is as follows:
wherein, betaiIs the weight value, T, of the ith time lag compensation submoduleiThe time constant of the ith time lag compensation submodule is taken as the time constant of the ith time lag compensation submodule; t isiThe value of (a) is between the maximum communication time lag and the minimum communication time lag of the system, so as to obtain better time lag compensation effect;
(13) calculating a transfer function ADC(s) of the self-adaptive time-lag compensator according to the transfer function of each time-lag compensation submodule and the corresponding weight; the transfer function ADC(s) of the adaptive skew compensator is calculated as follows:
according to the expression
It can be seen that after the compensation of the adaptive skew compensator, the phase lag of the wide-area measurement signal is only equal to the fixed time constant TcThe method is related and independent of the time lag tau, so that corresponding compensation can be made for different communication time lags;
(14) according to the expression
Obtaining a lag phase of the wide area measurement signal, which is irrelevant to the communication time lag, and compensating;
furthermore, in step (14), according to the frequency range of the system low-frequency oscillation, the lag phase in the frequency range is compensated by the lead-lag element.
Further, the step (2) comprises the following steps:
(21) amplifying the signal x (t) by k1Multiplying to obtain a first path of signal;
(22) amplifying the signal x (t) by k2Multiplying to obtain an intermediate signal x' (t);
(23) phase shifting the intermediate signal x' (t) to obtain a second path of signal;
(24) outputting a signal vector formed by the first path of signal and the second path of signal as a parallel phase-shifting signal X (t);
further, the mathematical expression of the phase shift in step (23) is
Wherein, TfThe filter constant is used for preventing the differential link from amplifying high-frequency noise to influence the control effect, and the value range of the filter constant is 0.01-0.05;
further, k is1And k2The normalization coefficient is used for ensuring that the amplitude ranges of the parallel first path of signals and the second path of signals are consistent.
Further, the step (3) comprises the following steps:
(31) setting parameters of a GrHDP neural network;
(32) randomly setting an initial weight of the GrHDP neural network, and performing off-line training on the GrHDP neural network by using the initial weight; taking the trained neural network weight as an initial weight for online learning;
(33) taking the parallel phase-shift signal X (t) as the input of a GrHDP neural network, taking the control signal u (t) as the output of the GrHDP neural network, and carrying out online application on the GrHDP neural network to obtain an optimized control signal; wherein the control signal u (t) includes an additional active power command value Δ Pref(t) and an additional reactive power command value DeltaQref(t) adding the active power command value Δ Pref(t) for compensating for oscillations in active power, adding a reactive power command value DeltaQref(t) for compensating oscillations of reactive power;
further, the parameters of the GrHDP neural network set in step (31) include: the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes, the learning rate, the upper limit of iteration times, the error tolerance and the weight range of the execution network; evaluating the number of nodes of an input layer, the number of nodes of a hidden layer, the number of nodes of an output layer, the learning rate, the upper limit of iteration times, the error tolerance and the weight range of the network; the number of nodes of an input layer, the number of nodes of a hidden layer, the number of nodes of an output layer, the learning rate, the upper limit of iteration times, the error tolerance and the weight range of the target network.
Generally, by the above technical solution conceived by the present invention, the following beneficial effects can be obtained:
(1) the method comprises the following steps that a wide-area measurement signal is used as an input, calculation from a parallel phase-shifting signal X (t) to a control signal u (t) is realized by a GrHDP unit based on a self-adaptive dynamic programming algorithm, and adaptive control can be performed on the current operation condition of a system without constructing a mathematical model of the system, so that the low-frequency oscillation of the system can be effectively inhibited through wide-area damping control under different operation conditions and fault disturbance of the system;
(2) the adaptive time lag compensator can effectively compensate the communication time lag of the wide area measurement signal, so that the adaptive wide area damping controller can keep good low-frequency oscillation suppression capability under different communication time lags, and the transient stability of the system is improved;
(3) the output control signal u (t) includes an additional active power command value Δ Pref(t) and an additional reactive power command value DeltaQref(t), an active control loop and a reactive control loop of the power system can be adjusted simultaneously, the damping ratio of a controlled module of the system is obviously improved, and the transient stability of the system is improved;
(4) parallel input signals are provided for the GrHDP neural network through phase shifting, so that the execution network can flexibly perform phase compensation on the input signals through weight adjustment, and the self-adaptive wide-area damping controller can better inhibit low-frequency oscillation of the system.
Drawings
FIG. 1 is a schematic diagram of an equivalent simplified model structure of a two-end alternating current system containing Yubei back-to-back flexible direct current;
FIG. 2 is a block diagram of an adaptive wide area damping controller provided in an embodiment of the present invention;
FIG. 3 is a flowchart of a control method according to an embodiment of the present invention;
fig. 4 is a transient characteristic curve diagram of the Hubei equivalent power grid, (a) is a transient characteristic curve diagram of the Hubei equivalent power grid under the situation I, and (b) is a transient characteristic curve diagram of the Hubei equivalent power grid under the situation II;
FIG. 5 is a diagram of an adaptive wide area damping controller internal variable curve under scenario II; (a) error E for implementing networkaA change curve; (b) is a variation curve of an external reinforcement learning function r (t); (c)is a variation curve of an internal reinforcement learning function S (t); (d) is the variation curve of the cost function J (t); (e) for performing weighting W from input layer to hidden layer of networka(1)A change curve; (f) for performing weights W from hidden layer to output layer of networka(2)A change curve;
FIG. 6 is a schematic diagram of the control performance of the adaptive wide-area damping controller provided in the embodiment of the present invention at a fixed time lag; (a) the fixed time lag is 100 ms; (b) the fixed time lag is 150 ms;
fig. 7 is a schematic diagram of the control performance of the adaptive controller provided in the embodiment of the present invention under a small range of random time lag;
fig. 8 is a graph illustrating weight variation of a time lag compensation submodule in the adaptive time lag compensator according to the embodiment of the present invention; (a) the random time lag range is 100 +/-20 ms; (b) the random time lag range is 100 +/-40 ms;
FIG. 9 is a schematic diagram of the control performance of the adaptive wide-area damping controller provided in the embodiment of the present invention under a large range of random time lags;
FIG. 10 is a diagram illustrating a random skew variation curve and a weight variation of a skew compensation sub-module in an adaptive skew compensator; (a) the change curve of the random time lag of the system is shown; (b) the weight value change curve of the time lag compensation submodule is obtained.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The two-end alternating current system containing the Yubei back-to-back flexible direct current shown in the figure 1 comprises 4 current conversion units of +/-420 kV/1250MW, and the total transmission power reaches 5000 MW; the inversion side alternating current system is an equivalent model of a power grid in Hubei province and comprises 5 equivalent generators and 4 equivalent loads; the provincial section tide of the power grid in the north of the lake and the adjacent provinces of the power grids in the south of the river, the west of the river and the south of the lake is equivalent to load, and 4 returns to the power grid in the north of the lakeThe China east power grid and the three gorges direct current transmission project from 1 back to the Guangdong are also equivalent to loads; g5、G6、G7、G8And G9The power generation system is characterized by comprising 5 equivalent generators in a power grid in the northlake, and lines 11-113 are power transmission lines in the power grid in the northlake; the rectification side alternating current system is an equivalent model of a southwest power grid (Sichuan and Chongqing power grids), and comprises 4 generators and 3 loads; the direct current engineering from 3 times in the Sichuan power grid to the east China power grid is equivalent according to the load; g1、G2、G3And G4The power generation system comprises 4 equivalent generators in a southwest power grid respectively, and lines 1-5 are power transmission lines in the southwest power grid; the minimum value of the communication time lag of the actual power grid is 50ms, and the maximum value is 500 ms; equivalent simplified system models of a southwest power grid and a Hubei power grid of a back-to-back flexible direct current transmission system containing Yubei shown in figure 1 are constructed in MATLAB/Simulink to serve as a test system of the embodiment of the invention.
The self-adaptive dynamic programming algorithm can learn an optimal control strategy through real-time interaction with the system; the GrHDP neural network realizes an adaptive dynamic programming algorithm based on the neural network, and comprises the following steps: an execution network, a target network and an evaluation network; the execution network is used for generating an output signal according to the input signal; the evaluation network is a function approximator, the output of the evaluation network is a cost function J (t), the cost function J (t) is used for evaluating the quality of the current output signal and guiding the execution network to carry out weight correction, so that the output signal is optimized; the target network is used for automatically generating the internal enhanced signal S (t) to replace the external enhanced signal r (t), so that the mapping relation between the input signal and the output signal is better reflected, and the cost function J (t) can better evaluate the quality of the output signal.
Fig. 2 illustrates an adaptive wide-area damping controller according to an embodiment of the present invention, including: the device comprises a self-adaptive time lag compensator, a phase shift unit and a GrHDP unit; the input end of the self-adaptive time lag compensator is used for receiving wide-area measurement signals
Adaptive skew compensator for wide area measurement signals to an electrical power system
Carrying out self-adaptive time lag compensation to obtain a signal x (t); the input end of the phase shifting unit is connected to the output end of the self-adaptive time lag compensator, and the phase shifting unit is used for amplifying and phase-shifting the signal X (t) to obtain a parallel phase-shifted signal X (t); the input end of the GrHDP unit is connected to the output end of the phase shifting unit, and the GrHDP unit is used for obtaining a control signal u (t) adaptive to the current operation condition of the power system according to the parallel phase shifting signal X (t) so as to realize adaptive compensation of active power and reactive power of the power system and effectively inhibit low-frequency oscillation of the power system;
the adaptive compensator comprises n time lag compensation submodules, the transfer function of the adaptive compensator is the weighted sum of the transfer functions of the n time lag compensation submodules, and the weight value of each time lag compensation submodule and the wide-area measurement signal
The communication time lag tau is related, so that the self-adaptive time lag compensator can correspondingly compensate different communication time lags, and the communication time lag measurement signal in a wide area can be effectively eliminated
The lag phase introduced therein; in this embodiment, n takes the value of 5;
the phase shifting unit comprises a first amplifier, a second amplifier and a phase shifter; the first amplifier is used for amplifying the signal x (t) by k1Multiplying to obtain a first path of signal; a second amplifier for amplifying the signal x (t) by k2Multiplying to obtain an intermediate signal x' (t); the input end of the phase shifter is connected to the output end of the second amplifier, and the phase shifter is used for shifting the phase of the intermediate signal x' (t) to obtain a second path of signal; the parallel phase shift signal X (t) comprises a first path of signal and a second path of signal; the mathematical expression of the phase shifter is
Wherein, TfAs filter constants for preventing the differential element from amplifying high-frequency noiseThe control effect is influenced, and the value range is 0.01-0.05; k is a radical of1And k2The normalization coefficient is used for ensuring that the amplitude ranges of the parallel first path of signals and the second path of signals are consistent;
the GrHDP unit realizes the calculation from the parallel phase-shifting signal X (t) to the control signal u (t) based on the self-adaptive dynamic programming algorithm; the control signal u (t) includes an additional active power command value Δ Pref(t) and an additional reactive power command value DeltaQref(t); additional active power command value Δ Pref(t) for compensating for oscillations in active power, adding a reactive power command value DeltaQref(t) for compensating oscillations of the reactive power.
Wide area measurement signal
The method is characterized in that the method is acquired by a Wide Area Measurement System (WAMS), and specific signal components are determined according to the control requirements of a power system.
Fig. 3 shows a control method based on the adaptive wide-area damping controller shown in fig. 2, which includes the following steps:
(1) for wide area measurement signal
Carrying out self-adaptive time lag compensation to obtain a signal x (t); the method specifically comprises the following steps:
(11) computing wide area control signals
Time-lag transfer function G ofd(s); and transfer function G is transformed by using second-order Pade approximationd(s) to obtain a transfer function GD(s), transfer function GDThe calculation formula of(s) is as follows:
wherein tau is a wide-area measurement signal
Communication skew of (2);
(12) calculating a transfer function and a corresponding weight of each time-lag compensation submodule; the transfer function calculation formula of each time lag compensation submodule is as follows:
wherein, TcIs a time constant associated with the steady state behavior of the system; taking into account the steady-state behavior of the system, TcThe value of (a) is 0.02 s;
the weight calculation formula of the time lag compensation submodule is as follows:
wherein, betaiIs the weight value, T, of the ith time lag compensation submoduleiThe time constant of the ith time lag compensation submodule is taken as the time constant of the ith time lag compensation submodule; t isiThe value of (1) is between the maximum communication time lag and the minimum communication time lag of the system so as to obtain a better time lag compensation effect; considering that the minimum value of the communication time lag of the actual power grid is 50ms and the maximum value is 500ms, setting T1＝0.1s，T2＝0.2s，T3＝0.3s，T4＝0.4s，T5＝0.5s；
(13) Calculating a transfer function ADC(s) of the self-adaptive time-lag compensator according to the transfer function of each time-lag compensation submodule and the corresponding weight; the transfer function ADC(s) of the adaptive skew compensator is calculated as follows:
according to the expression
It can be seen that after the compensation of the adaptive skew compensator, the phase lag of the wide-area measurement signal is only equal to the fixed time constant TcThe method is related and independent of the time lag tau, so that corresponding compensation can be made for different communication time lags;
(14) according to the expression
Obtaining phase lag of the wide area measurement signal, which is irrelevant to communication time lag, and compensating the phase lag in a frequency range through a lead-lag link according to the frequency range of low-frequency oscillation of the system;
(2) amplifying and phase-shifting the signal X (t) to obtain a parallel phase-shifted signal X (t); the method specifically comprises the following steps:
(21) amplifying the signal x (t) by k1Multiplying to obtain a first path of signal;
(22) amplifying the signal x (t) by k2Multiplying to obtain an intermediate signal x' (t); k is a radical of1And k2The normalization coefficient is used for ensuring that the amplitude ranges of the parallel first path of signals and the second path of signals are consistent;
(23) phase shifting the intermediate signal x' (t) to obtain a second path of signal; the mathematical expression for the phase shift is
Wherein, TfThe filter constant is used for preventing the differential link from amplifying high-frequency noise to influence the control effect, and the value range of the filter constant is 0.01-0.05;
(24) outputting a signal vector formed by the first path of signal and the second path of signal as a parallel phase-shifting signal X (t);
(3) obtaining a control signal u (t) adaptive to the current operation environment of the power grid according to the parallel phase-shift signal X (t) so as to realize the adaptive compensation of the active power and the reactive power of the power grid, thereby effectively inhibiting the low-frequency oscillation of the power system; the method specifically comprises the following steps:
(31) the parameters of the GrHDP neural network are set, and the specific settings are shown in table 1:
TABLE 1 parameter settings for GrHDP neural networks
| Executive network | Evaluation network | Target network | |
| Number of nodes of |
2 | 5 | 4 |
| Number of hidden |
3 | 3 | 3 |
| Number of |
2 | 1 | 2 |
| Learning rate | 0.02 | 0.01 | 0.01 |
| Upper limit of iteration number | 50 | 50 | 50 |
| Margin of error | 1e-8 | 1e-8 | 1e-8 |
| Weight range | ±5 | ±5 | ±5 |
(32) Randomly setting an initial weight of the GrHDP neural network, and performing off-line training on the GrHDP neural network by using the initial weight; taking the trained neural network weight as an initial weight for online learning;
(33) taking the parallel phase-shift signal X (t) as the input of a GrHDP neural network, taking the control signal u (t) as the output of the GrHDP neural network, and carrying out online application on the GrHDP neural network to obtain an optimized control signal; wherein the control signal u (t) includes an additional active power command value Δ Pref(t) and an additional reactive power command value DeltaQref(t) adding the active power command value Δ Pref(t) for compensating for oscillations in active power, adding a reactive power command value DeltaQref(t) for compensating oscillations of the reactive power.
Setting a first example for verifying the adaptability of the adaptive wide area damping controller to the system working condition; a second example is provided for verifying the effect of the adaptive skew compensator in compensating for system signal communication skews.
In the first example, the variation working condition deviating from the typical operation working condition of the WADC is obtained by adjusting the output and active load of the generator in the Hubei equivalent power grid; under the changing working condition, the linear modal analysis is carried out on the Hubei equivalent power grid when the WADC is not put into the power grid, the damping ratio of the mode 1 is-1.88 percent, and a negative damping state is presented; in order to verify the adaptability of the A-WADC to the system working condition, two sets of scenarios are respectively set:
scenario I: under a typical operation condition, at 1 second, in the Hubei equivalent power grid, a permanent three-phase short circuit fault occurs at a position, close to a bus 18, of one circuit in double-circuit power transmission lines 19-111, and the fault circuit is cut off within 1.1 second;
scenario II: under the condition of changing operation conditions, instantaneous three-phase short-circuit faults occur at the positions, close to the bus 18, of the power transmission lines 17-18 in the Hubei equivalent power grid at 1 second, the fault lines are cut off at 1.1 second, and reclosing is successful at 1.8 seconds.
Comparing transient response characteristics of the system after system failure when different damping controllers are put into operation respectively, and fig. 4 shows a relative power angle change curve of the generator G5 and the generator G9 under two situations; fig. 5 shows an internal variable curve of a-WADC under scenario II, which includes an execution network error Ea, an external reinforcement learning function r (t), an internal reinforcement learning function S (t), a cost function J (t), a weight Wa (1) from an execution network input layer to a hidden layer, and a weight Wa (2) from the execution network hidden layer to an output layer.
As shown in fig. 4, under the typical condition of the scenario I, when a conventional wide-area damping controller (C-WADC) and a trained adaptive wide-area damping controller (a-WADC) are put into operation, the low-frequency oscillation of the system can be quickly settled, and the control performance of the two controllers is substantially the same. Under the condition that the scene II changes the operation condition, when no damping controller is put into the system, the system presents the characteristic of amplified oscillation. When C-WADC or A-WADC is put into use, the oscillation of the system can be effectively subsided, and the control effect of the A-WADC is obviously better than that of the C-WADC. The reason is that the control parameters of the C-WADC designed based on the typical operation working condition cannot change along with the change of the system operation working condition, and when the system deviates from the typical operation working condition, the control performance of the C-WADC is reduced; the A-WADC can adapt to the changing operation condition of the system by updating the weight of the neural network on line, thereby keeping better oscillation suppression effect.
When the system is disturbed and the running state is changed, the input signal of the A-WADC generates low-frequency oscillation as shown in figure 4; at this time, as shown in fig. 5, the external reinforcement learning function r (t), the internal reinforcement learning function S (t), and the cost function J (t) also fluctuate accordingly, so that the error Ea of the execution network exceeds the error tolerance value, and the execution network starts to correct the weight Wa (1) and the weight Wa (2); in the process, the output control signal of the A-WADC is optimized, the weight correction of the execution network is basically finished within about 4 seconds, the A-WADC adapts to the new system running state again, and meanwhile, the low-frequency oscillation of the system is basically subsided. Simulation results verify that the A-WADC can realize online self-learning through updating of weights of the neural network, further adapt to changes of system operation conditions, and can keep good low-frequency oscillation suppression capability of the system under different conditions and different faults.
In the second example, in order to verify the compensation effect of the ADC module on the system communication skew, the following three scenarios are set respectively:
scenario III: fixed time lag
The communication skew of the system is set to 100ms and 150ms, respectively. The operation condition and the fault setting of the system are the same as those of the scenario II, and the transient response characteristics of the system under different fixed communication time lags after the system is in fault are compared when different damping controllers are put into the system. Fig. 6 shows the relative power angle curves of the generators G5 and G9 at different fixed time lag levels.
As shown in fig. 6(a), when the time lag is 100ms, the a-WADC with the ADC can suppress the low frequency oscillation of the system more rapidly than the a-WADC without the ADC, and when the C-WADC is applied, the low frequency oscillation of the system decays very slowly; as shown in fig. 6(b), when the time lag is 150ms, when the a-WADC including the ADC is input, the low-frequency oscillation of the system can still be rapidly subsided, and when the a-WADC and the C-WADC not including the ADC are input, the system respectively generates amplified oscillation and constant-amplitude oscillation, which shows that the communication time lag at this time seriously affects the weight correction process of the a-WADC and the control effect of the C-WADC.
Therefore, as the communication time lag is increased, the effect of the damping controller without the time lag compensation capability on inhibiting the low-frequency oscillation of the system is reduced; when the communication time lag increases to a certain degree, the transient stability of the system is even deteriorated; meanwhile, the simulation result shown in fig. 6 also shows that the adaptive skew compensator can well compensate different fixed communication skews, so that the a-WADC including the ADC maintains a good capability of suppressing the low-frequency oscillation of the system, and the transient stability of the system is improved.
Scenario IV: small range random communication time lag
Setting the random range of the communication time lag of the system as 100 plus or minus 0ms (fixed time lag), 100 plus or minus 20ms, 100 plus or minus 40ms and 100 plus or minus 60ms respectively; the operation condition and the fault setting of the system are the same as those in the scenario II, at the moment, the back-to-back flexible direct system is put into the A-WADC containing the ADC, and after the system is disturbed, the transient response characteristics of the system under different random communication time lags are compared; FIG. 7 shows the relative power angle curves of generators G5 and G9 at different random communication skew levels; fig. 8 is a graph showing the variation of weights of the skew compensation sub-modules in the ADC under different random communication skew levels.
As shown in fig. 7, the control effect under the fixed communication time lag is better than the control effect with the random communication time lag under different random ranges, and the effect of the controller shows a decreasing trend as the random range of the communication time lag increases; in addition, under different random communication time lag ranges, the A-WADC containing the ADC can quickly inhibit the low-frequency oscillation of the system, which shows that the ADC has better capability of compensating the small-range random time lag.
As shown in fig. 8(a), the weight β 1 fluctuates around 1, the weights of other SDCs fluctuate around 0, and the fluctuation amplitude appears β2>β3>β4>β5The rule of (2); as the random range of time lag is increased, the fluctuation amplitude of 5 weights is also increased; the ADC realizes the compensation of random time lag just by continuously adjusting the weight of the SDC.
Scene V: large range random communication skew
Setting the random range of communication time lag of the system to be 50-500 ms, setting the operation working condition and the fault setting of the system to be the same as the setting of a scene II, and comparing the transient response characteristics of the system after the fault when different damping controllers are put into use; FIG. 9 shows the relative power angle curves of generators G5 and G9 under the control of different damping controllers; fig. 10 is a graph showing the random communication time lag and the weight variation of the time lag compensation submodule in the ADC.
As shown in fig. 9, under the influence of large-scale random time lag, when C-WADC and a-WADC without ADC are used, the system has amplified oscillation, and compared with the transient characteristic of the system without WADC, it is shown that the transient characteristics of the system cannot be improved by the two controllers at this time; the A-WADC containing the ADC can still quickly inhibit the low-frequency oscillation of the system, which shows that the ADC has better capability of compensating large-range random time lag.
Fig. 10(a) is a graph showing a change in random communication time lag; fig. 10(b) is a weight variation curve of the skew compensation submodule in the corresponding ADC; as shown in fig. 10, when the random skew fluctuates in a large range, the ADC can still better compensate the skew of the wide-area control signal by adjusting the weight of each SDC on line, and maintain the control performance of the a-WADC.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (8)
1. An adaptive wide area damping controller, comprising: the device comprises a self-adaptive time lag compensator, a phase shift unit and a GrHDP unit;
the input end of the self-adaptive time lag compensator is used for receiving wide area measurement signals
The adaptive time lag compensator is used for wide area measurement signals of the power system
Carrying out self-adaptive time lag compensation to obtain a signal x (t);
the input end of the phase shift unit is connected to the output end of the self-adaptive time lag compensator, and the phase shift unit is used for amplifying and phase-shifting a signal X (t) to obtain a parallel phase-shifted signal X (t);
the input end of the GrHDP unit is connected to the output end of the phase shift unit, and the GrHDP unit is used for obtaining a control signal u (t) adaptive to the current operation condition of the power system according to the parallel phase shift signal X (t) so as to realize the power supplySelf-adaptive compensation of active power oscillation and reactive power oscillation of the system is carried out, so that low-frequency oscillation of the power system is effectively inhibited; the control signal u (t) comprises an additional active power command value Δ Pref(t) and an additional reactive power command value DeltaQref(t); the additional active power command value Δ Pref(t) for compensating for oscillations in active power, said additional reactive power command value Δ Qref(t) for compensating oscillations of reactive power;
the electric power system is a two-end alternating current system containing Yuhui back-to-back flexible direct current, and the time lag compensated by the self-adaptive time lag compensator is
TcIs a time constant associated with the steady state behavior of the system.
2. The adaptive wide area damping controller of claim 1, wherein the adaptive skew compensator includes n skew compensation submodules; the transfer function of the self-adaptive time lag compensator is the weighted sum of the transfer functions of the n time lag compensation submodules, and the weight value of each time lag compensation submodule and the wide-area measurement signal
The self-adaptive time lag compensator can correspondingly compensate different communication time lags so as to effectively eliminate communication time lag measurement signals in a wide area
The lag phase introduced in.
3. The adaptive wide area damping controller of claim 1, wherein the phase shifting unit comprises a first amplifier, a second amplifier, and a phase shifter;
the first amplifier is used for amplifying the signal x (t) by k1Multiplying to obtain a first path of signal;
the second amplifier is used for amplifying the signal x (t) by k2Multiplying to obtain an intermediate signal x' (t);
the input end of the phase shifter is connected to the output end of the second amplifier, and the phase shifter is used for shifting the phase of the intermediate signal x' (t) to obtain a second path of signal;
the parallel phase-shift signal X (t) comprises the first path of signal and the second path of signal;
wherein k is1And k2The normalization coefficient is used for ensuring that the amplitude ranges of the parallel first path of signals and the second path of signals are consistent.
4. A control method based on an adaptive wide area damping controller according to any of claims 1-3, characterized by the steps of:
(1) for wide area measurement signal
Carrying out self-adaptive time lag compensation to obtain a signal x (t);
(2) amplifying and phase-shifting the signal X (t) to obtain a parallel phase-shifted signal X (t);
(3) and obtaining a control signal u (t) adaptive to the current operating environment of the power grid according to the parallel phase-shifting signal X (t) so as to realize the adaptive compensation of the active power oscillation and the reactive power oscillation of the power grid, thereby effectively inhibiting the low-frequency oscillation of the power system.
5. The control method according to claim 4, wherein the step (1) includes the steps of:
(11) calculating the wide area measurement signal
Time-lag transfer function G ofd(s); and using second-order Pade approximate transformation to transfer function Gd(s) to obtain a transfer function GD(s), the transfer function GDFormula for calculating(s)The following were used:
wherein τ is the wide-area measurement signal
Communication skew of (2);
(12) calculating a transfer function and a corresponding weight of each time-lag compensation submodule; the transfer function calculation formula of each time lag compensation submodule is as follows:
wherein, TcIs a time constant associated with the steady state behavior of the system;
the weight calculation formula of the time lag compensation submodule is as follows:
wherein, betaiIs the weight value, T, of the ith time lag compensation submoduleiThe time constant of the ith time lag compensation submodule is taken as the time constant of the ith time lag compensation submodule; t isiThe value of (a) is between the maximum communication time lag and the minimum communication time lag of the system, so as to obtain better time lag compensation effect;
(13) calculating a transfer function ADC(s) of the self-adaptive time-lag compensator according to the transfer function of each time-lag compensation submodule and the corresponding weight; the transfer function ADC(s) of the adaptive skew compensator is calculated as follows:
(14) according to the expression
And obtaining a lag phase of the wide area measurement signal, which is irrelevant to the communication time lag, and compensating.
6. The control method according to claim 4, wherein the step (2) specifically includes the steps of:
(21) amplifying the signal x (t) by k1Multiplying to obtain a first path of signal;
(22) amplifying the signal x (t) by k2Multiplying to obtain an intermediate signal x' (t);
(23) shifting the phase of the intermediate signal x' (t) to obtain a second path of signal;
(24) outputting a signal vector formed by the first path of signal and the second path of signal as a parallel phase-shifting signal X (t);
wherein k is1And k2The normalization coefficient is used for ensuring that the amplitude ranges of the parallel first path of signals and the second path of signals are consistent.
7. The control method according to claim 4, wherein the step (3) specifically includes the steps of:
(31) setting parameters of a GrHDP neural network;
(32) randomly setting an initial weight of the GrHDP neural network, and performing off-line training on the GrHDP neural network by using the initial weight; taking the trained neural network weight as an initial weight for online learning;
(33) taking the parallel phase-shift signal X (t) as the input of a GrHDP neural network, taking the control signal u (t) as the output of the GrHDP neural network, and carrying out online application on the GrHDP neural network to obtain an optimized control signal; wherein the control signal u (t) comprises an additional active power command value Δ Pref(t) and an additional reactive power command value DeltaQref(t) the additional active power command value Δ Pref(t) for compensating for oscillations in active power, said additional reactive power command value Δ Qref(t) for compensating oscillations of the reactive power.
8. The control method according to claim 7, wherein the parameters of the neural network set in the step (31) include: the number of input layer nodes, the number of hidden layer nodes, the number of output layer nodes, the learning rate, the upper limit of iteration times, the error tolerance and the weight range of the execution network; evaluating the number of nodes of an input layer, the number of nodes of a hidden layer, the number of nodes of an output layer, the learning rate, the upper limit of iteration times, the error tolerance and the weight range of the network; the number of nodes of an input layer, the number of nodes of a hidden layer, the number of nodes of an output layer, the learning rate, the upper limit of iteration times, the error tolerance and the weight range of the target network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810126049.XA CN108365615B (en) | 2018-02-08 | 2018-02-08 | Self-adaptive wide area damping controller and control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810126049.XA CN108365615B (en) | 2018-02-08 | 2018-02-08 | Self-adaptive wide area damping controller and control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108365615A CN108365615A (en) | 2018-08-03 |
| CN108365615B true CN108365615B (en) | 2021-02-09 |
Family
ID=63005069
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810126049.XA Active CN108365615B (en) | 2018-02-08 | 2018-02-08 | Self-adaptive wide area damping controller and control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108365615B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109785289B (en) * | 2018-12-18 | 2021-07-20 | 中国科学院深圳先进技术研究院 | Transmission line defect detection method and system and electronic equipment |
| CN109725534A (en) * | 2018-12-29 | 2019-05-07 | 云南电网有限责任公司电力科学研究院 | The adaptive dynamic programming method of STATCOM controller based on MMC |
| CN109742773B (en) * | 2019-01-29 | 2020-07-28 | 华中科技大学 | Self-adaptive wide area damping controller |
| CN112350343A (en) * | 2019-08-09 | 2021-02-09 | 国家电网公司华东分部 | Controllable phase shifter damping control method taking power as input quantity |
| CN111384717B (en) * | 2020-01-15 | 2022-02-18 | 华中科技大学 | Adaptive damping control method and system for resisting false data injection attack |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101465550B (en) * | 2007-12-21 | 2011-01-26 | 清华大学 | Sdelayed time processing and compensating system for wide domain damped control of electric power |
| CN102624013B (en) * | 2012-04-06 | 2015-01-07 | 湖北省电力公司 | Phase compensation principle-based design method for energy storage damping controller |
| CN105117535B (en) * | 2015-08-12 | 2018-05-08 | 浙江工业大学 | Suitable for the electrical power system wide-area PID damping controller design methods of stochastic Time-Delay |
-
2018
- 2018-02-08 CN CN201810126049.XA patent/CN108365615B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108365615A (en) | 2018-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108365615B (en) | Self-adaptive wide area damping controller and control method | |
| Shen et al. | Adaptive supplementary damping control of VSC-HVDC for interarea oscillation using GrHDP | |
| Renedo et al. | Reactive-power coordination in VSC-HVDC multi-terminal systems for transient stability improvement | |
| Tiwari et al. | Neural network predictive control of UPFC for improving transient stability performance of power system | |
| Surinkaew et al. | Inter-area oscillation damping control design considering impact of variable latencies | |
| CN109742773B (en) | Self-adaptive wide area damping controller | |
| Yao et al. | Adaptive power oscillation damping controller of superconducting magnetic energy storage device for interarea oscillations in power system | |
| CN115102149A (en) | Overcurrent suppression system and method for network type converter | |
| Dehkordi et al. | Voltage and frequency consensusability of autonomous microgrids over fading channels | |
| Ibrahim et al. | Performance assessment of bacterial foraging based power system stabilizer in multi-machine power system | |
| Preece et al. | Damping of inter-area oscillations using WAMS based supplementary controller installed at VSC based HVDC line | |
| Sanz et al. | Coordinated corrective control for transient stability enhancement in future Great Britain transmission system | |
| Marei et al. | An intelligent control for the DG interface to mitigate voltage flicker | |
| Roy et al. | Hybrid robust adaptive backstepping sliding mode controller design for mitigating SSR in series-compensated DFIG-based wind generation systems | |
| Wen et al. | Cascaded sliding-mode observer and its applications in output feedback control part II: Adaptive output feedback control | |
| Lazrak et al. | An improved control strategy for DFIG wind turbine to ride-through voltage dips | |
| Liu et al. | Markov-Based Stochastic Stabilization Control for MMC-HVDC Systems with Inertia Supporting Under Random Disturbances | |
| Bhadu et al. | Design and analysis of noise extenuation techniques in modern LFC system | |
| Hossen et al. | Tunicate swarm algorithm for power system stability enhancement in a SMIB-UPFC network | |
| Hashemi et al. | Simultaneous coordinated tuning of UPFC and multi-input PSS for damping of power system oscillations | |
| Touhami et al. | Performance evaluation of fuzzy-logic controller applied to a UPFC transmission system | |
| Muppoori et al. | Critical Assessment and Comparative Study of PID and ADRC Approaches Applied to AGC in Multi-Source Single Area Power System | |
| Sharma et al. | Enhanced modular multi-level converter-based static synchronous compensator with social ski-driver optimized adaptive neuro-fuzzy inference system controller for power quality enhancement | |
| Tapia et al. | Power systems neural voltage control by a STATCOM | |
| Naqvi et al. | Grid integration of a three phase multifunctional SECS using Lorentzian adaptive filter based control with impulsive disturbance rejection capability |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |