Method, system and medium for detecting sub-harmonic oscillation based on tunable notch filter

Abstract

The application provides a method and system for detecting sub-harmonic oscillation based on a tunable wave trap, and a medium. A direct current digital filter and a tunable digital wave trap are cascaded to obtain a single-frequency oscillation signal detector. A wave trap angle frequency of the tunable wave trap is used as an independent variable, and an energy function is constructed based on a to-be-detected oscillation signal and an output of the single-frequency oscillation signal detector. A gradient descent method is used to iteratively update the independent variable in a negative gradient direction of the energy function with a fixed step length, to search for a minimum value of the energy function. When a gradient of the energy function with respect to the independent variable is zero, the corresponding angle frequency is determined as an angle frequency of the to-be-detected oscillation signal. The implementation process is simple and easy to implement, the calculation amount is small, and the real-time and accurate detection of the sub-harmonic oscillation signal can be realized.

Description

Technical Field

[0001] This invention relates to the field of signal frequency detection technology, specifically to a method, system, and medium for detecting sub-supersynchronous oscillations based on a tunable notch filter. Background Technology

[0002] With the large-scale integration of new energy sources such as wind power and photovoltaics into the power system, the application of power electronic equipment in the grid connection process is becoming increasingly common. The interaction between power electronic converters and grid impedance can easily generate various non-power frequency oscillations in the power grid. Oscillations in traditional power systems are mostly concentrated in the subsynchronous range, while in the scenario of new energy grid connection, the oscillation frequency range is further expanded, including not only subsynchronous oscillations, but also supersynchronous oscillations above the power frequency, and may even extend to high-frequency harmonic oscillations in the kilohertz range.

[0003] Subsynchronous / supersynchronous oscillations typically occur between several Hz and twice the power frequency. They possess significant energy and can cause large, periodic fluctuations in current and voltage, leading to unstable power transmission, transformer mechanical vibration, and capacitor overheating damage, seriously threatening the safe and stable operation of the power system. Therefore, real-time and accurate detection of subsynchronous / supersynchronous oscillation signals is a prerequisite for subsequent analysis and suppression measures.

[0004] Currently widely used oscillation signal detection methods based on Fast Fourier Transform (FFT) and Discrete Fourier Transform (DFT) can simultaneously detect the power frequency, harmonics, and subsynchronous / supersynchronous harmonics of the power grid. However, DFT and FFT are difficult to use for synchronous sampling during subsynchronous / supersynchronous oscillations, leading to significant errors in frequency detection due to spectral leakage and the picket fence effect. Furthermore, they are computationally intensive and have poor real-time performance. Phase-locked loops (PLLs) are primarily suitable for real-time detection of the power grid frequency, but not for detecting the frequency of time-varying subsynchronous / supersynchronous harmonic oscillations over a wide frequency range. The step response time of a PLL for detecting the power grid frequency is typically 2 to 4 times the period of the power frequency voltage signal. For example, for a 50Hz power frequency signal, the step response time of a PLL is generally 120ms to 200ms.

[0005] Therefore, there is an urgent need for a real-time detection method that has low computational cost, fast detection speed, and can cover the sub / supersynchronous frequency range. Summary of the Invention

[0006] This invention proposes a sub-supersynchronous oscillation detection method, system, and medium based on a tunable notch filter, to solve the technical problems of existing technologies such as large computational load, slow detection speed, and difficulty in covering the sub / supersynchronous frequency range.

[0007] To address the aforementioned technical problems, this invention provides a sub-supersynchronous oscillation detection method based on a tunable notch filter, comprising the following steps:

[0008] Step S1: Cascade a DC digital filter with a tunable digital notch filter to obtain a single-frequency oscillation signal detector;

[0009] Step S2: Using the notch angular frequency of the tunable notch filter as the independent variable, construct an energy function based on the oscillation signal to be measured and the output of the single-frequency oscillation signal detector;

[0010] Step S3: Using the gradient descent method, iteratively update the energy function along the negative gradient direction of the independent variable with a fixed step size to search for the minimum value of the energy function: when the gradient of the energy function with respect to the independent variable is zero, the corresponding angular frequency is determined as the angular frequency of the oscillation signal to be measured.

[0011] Preferably, the expression for the tunable digital notch filter is... for:

[0012] ;

[0013] In the formula, The sampling interval; The independent variable is the discretized system variable; This is the notch angular frequency.

[0014] Preferably, the expression of the DC digital filter for:

[0015] .

[0016] Preferably, the energy function The expression is:

[0017] ;

[0018] In the formula, The independent variable is the discretized system variable; This indicates the cumulative delay count for a single calculation; Indicates the current time; Indicates a single-frequency oscillation signal detector; This represents the value of the single-frequency oscillation signal detector at the s-th beat; The input single-frequency oscillation signal is the oscillation signal to be measured by the detector.

[0019] Preferably, the cumulative delay number for each single calculation is... Set the following conditions:

[0020] ;

[0021] In the formula, Indicates the sampling interval; Indicates the notch filter frequency.

[0022] Preferably, when the oscillation signal to be tested is a multi-frequency oscillation signal, segmented filtering preprocessing is performed first, and then the frequency of the oscillation components in different frequency bands is detected separately through step S3.

[0023] The present invention also provides a sub-supersynchronous oscillation detection system based on a tunable notch filter, implemented based on the above method, including a DC digital filter module, a tunable digital notch filter module, a cascaded detector module, an energy function calculation module, an iterative calculation module, and a frequency determination module;

[0024] The DC digital filter module is used to filter the input signal under test to eliminate DC bias.

[0025] A tunable digital notch filter module is used to perform notch filtering on the signal under test so that the output is zero when the notch filter frequency is the same as the frequency of the single-frequency sinusoidal signal under test.

[0026] The cascaded detector module, which is composed of the DC digital filter module and the tunable digital notch filter module connected in series, is used to output the detection result of a single-frequency oscillation signal.

[0027] An energy function calculation module is used to construct an energy function based on the output of the cascaded detector module, with the notch angular frequency of the tunable digital notch module as the independent variable.

[0028] The iterative calculation module is used to iteratively update the independent variable based on the energy function, using the gradient descent method and with a fixed step size along the negative gradient direction.

[0029] The frequency determination module is used to output the corresponding angular frequency as the frequency of the oscillation signal under test when the gradient of the energy function is zero.

[0030] The present invention also provides a computer-readable storage medium storing a program that, when executed by a processor, implements the above-described method.

[0031] The beneficial effects of this invention include at least the following: By constructing a tunable digital notch filter and combining it with a DC filter to form a single-frequency detector, and using an energy function as the core, this invention employs a gradient descent iterative method to achieve real-time search and determination of the oscillation frequency. This enables rapid and accurate detection of subsynchronous / supersynchronous oscillation signals in new energy grid-connected systems without the need for complex Fourier transform calculations. The overall response speed of this invention is significantly faster than that of the phase-locked loop method, and it overcomes the spectral leakage and picket-fence effects present in Fast Fourier Transform and Discrete Fourier Transform. It boasts high detection accuracy, low computational complexity, and is suitable for embedded device implementation. Furthermore, this invention can cover a wideband oscillation range from several Hz to twice the power frequency, and can achieve independent detection of multiple frequency oscillation components through segmented filtering and parameter adjustment, exhibiting excellent real-time performance, applicability, and engineering application value. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0033] Figure 2 This is a schematic diagram of the amplitude-frequency response of a tunable notch filter according to an embodiment of the present invention;

[0034] Figure 3 Different embodiments of the present invention M A schematic diagram of the energy function iteration under the given value;

[0035] Figure 4 This is a schematic diagram of the energy function iteration under different iteration step sizes in an embodiment of the present invention;

[0036] Figure 5 This is a schematic diagram illustrating the energy function iteration under different initial values ​​of the independent variable according to an embodiment of the present invention;

[0037] Figure 6 This is a schematic diagram of the frequency detection process of a parallel tunable notch / filter system for multi-frequency disturbance signals according to an embodiment of the present invention.

[0038] Figure 7 Different embodiments of the present invention M A schematic diagram of the signal frequency detection results at the specified value;

[0039] Figure 8 This is a schematic diagram of signal frequency detection results under different iteration step sizes in an embodiment of the present invention;

[0040] Figure 9 This is a schematic diagram of the signal frequency detection results at different test frequencies according to an embodiment of the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0042] like Figure 1 As shown, this embodiment of the invention provides a sub-supersynchronous oscillation detection method based on a tunable notch filter, comprising the following steps:

[0043] Step S1: Cascade a DC digital filter with a tunable digital notch filter to obtain a single-frequency oscillation signal detector.

[0044] For example, in this embodiment of the invention, for an angular frequency of The tunable digital notch filter constructed from the sinusoidal signal is as follows:

[0045] (1)

[0046] in The sampling interval is... The independent variable is the system after discretization. The notch angular frequency, , The notch filter frequency is given by the amplitude-frequency response of the notch filter as follows: Figure 2 As shown.

[0047] like Figure 2 As shown, when the angular frequency At that time, the amplitude of the tunable notch filter is:

[0048] (2)

[0049] Equation (2) shows that when the input angular frequency to the notch filter is When a sinusoidal signal is input, the output of the notch filter is zero, meaning that the input sinusoidal signal is completely filtered out.

[0050] To eliminate the DC bias in the secondary detection signal, the DC digital filter constructed in this embodiment of the invention is as follows:

[0051] (3)

[0052] Finally, by cascading the tunable digital notch filter with a DC digital filter, we obtain the tunable digital notch filter, which is also a single-frequency oscillation signal detector, as follows:

[0053] (4)

[0054] Step S2: Using the notch angular frequency of the tunable notch filter as the independent variable, construct an energy function based on the output of the oscillation signal under test and the single-frequency oscillation signal detector.

[0055] Specifically, for the tunable notch filter in equation (4), the energy function can be constructed as follows:

[0056] (5)

[0057] in To input the signal to be detected into the notch filter, Let ω be the independent variable, angular frequency. , Let be the expression for the single-frequency oscillation signal detector at the s-th beat. The accumulated delay is calculated in a single operation, and its magnitude is determined by... The initial state duration set initially.

[0058] Step S3: Using the gradient descent method, iteratively update the energy function along the negative gradient direction of the independent variable with a fixed step size to search for the minimum value of the energy function: when the gradient of the energy function with respect to the independent variable is zero, the corresponding angular frequency is determined as the angular frequency of the oscillation signal to be measured.

[0059] Specifically, the solution process is as follows: In equation (5), the current | |Status and The energy function output is calculated by summing the delay states, which improves the system's detection efficiency. The iterative diagram of the energy function under the value is shown below. Figure 3 As shown. In order to minimize potential oscillations while ensuring rapid detection, the embodiments of the present invention select... The following conditions must be met:

[0060] (6)

[0061] Then, the energy function is iteratively calculated using the gradient descent method. Its gradient at the current point is solved, and the search continues iteratively in the opposite direction at a predetermined step distance. The gradient expression of the energy function with respect to its independent variable at a certain point is as follows:

[0062] (7)

[0063] Among them, the tunable notch / filter right The partial derivative is written as follows:

[0064] (8)

[0065] The energy function is minimized by searching along the gradient in the opposite direction with a fixed step size. The search method is expressed as follows:

[0066] (9)

[0067] in as independent variable Preset initial value, when When =1, ; Let the search step size be _____. When the search step size is _____, the search step size is _____. When the value is large, the independent variable in two consecutive iterations When the difference is large, the iteration speed is faster, but the independent variable may directly exceed the actual frequency during the iteration process due to the large span, resulting in the detection results failing to converge. Conversely, a smaller span will lead to a slower search speed, failing to meet the speed requirements of frequency detection. The iterative diagram of the energy function under different iteration step sizes is shown in the figure. Figure 4 As shown.

[0068] Furthermore, in the initial value of the independent variable Under the same conditions, if the signal to be detected They are different, but use the same search step size. Since different signals to be detected correspond to different energy function gradients, their detection speeds will also differ. A schematic diagram of the energy function iteration under different initial values ​​of the independent variables is shown below. Figure 5 As shown.

[0069] When the gradient of the energy function with respect to its independent variable is zero, the corresponding frequency is the frequency of the disturbance signal to be measured, that is, the frequency of the detected disturbance signal.

[0070] For a signal to be detected containing multiple oscillation frequencies, this embodiment of the invention performs segmented filtering preprocessing on the signal to be detected, and then detects the frequencies of the oscillation components in different frequency bands using a frequency detection method based on a tunable notch filter; furthermore, different frequencies are used for the oscillation components in different frequency bands. and To improve detection efficiency. A frequency detection scheme for multi-frequency oscillation signals is shown in the diagram below. Figure 6 As shown.

[0071] different The signal frequency detection results at the value are as follows Figure 7 As shown, the frequency of the signal under test is 10Hz, and the initial frequency value is set to 15Hz. It can be seen that... The larger the value, the longer the frequency remains in its initial state, but the faster the frequency detection speed.

[0072] Signal frequency detection results under different iteration step sizes are as follows Figure 8As shown, it can be seen that, in general, the larger the iteration step size, the faster the frequency detection speed. However, if the iteration step size is too large, the independent variable will exceed the actual frequency during the iteration process, which will reduce the detection efficiency.

[0073] Signal detection results at different test frequencies are as follows Figure 9 As shown, the frequencies of the signals to be tested are 15Hz and 85Hz, representing subsynchronous oscillation and supersynchronous oscillation signals, respectively, with an initial frequency value of 50Hz for both. It can be seen that this detection method can achieve real-time and accurate detection of subsynchronous and supersynchronous oscillation disturbance signals, solving the problem that while SRF-PLLs can detect in real time, they cannot detect subsynchronous and supersynchronous oscillation signals and can only detect oscillation signal frequencies at the power frequency and its integer multiples. Furthermore, its detection speed is approximately 50 to 60 ms, about 1 to 2 times the period of the detected frequency signal, significantly faster than the frequency detection speed of PLLs, demonstrating better real-time performance.

[0074] The present invention also provides a sub-supersynchronous oscillation detection system based on a tunable notch filter, implemented based on the above method, including a DC digital filter module, a tunable digital notch filter module, a cascaded detector module, an energy function calculation module, an iterative calculation module, and a frequency determination module;

[0075] The DC digital filter module is used to filter the input signal under test to eliminate DC bias.

[0076] The tunable digital notch filter module is used to perform notch filtering on the signal under test, so that the output is zero when the notch filter frequency is the same as the frequency of the single-frequency sinusoidal signal under test.

[0077] The cascaded detector module, consisting of a DC digital filter module and a tunable digital notch filter module connected in series, is used to output the detection result of a single-frequency oscillation signal.

[0078] The energy function calculation module is used to construct an energy function based on the output of the cascaded detector module, with the notch angular frequency of the tunable digital notch module as the independent variable.

[0079] The iterative calculation module is used to iteratively update the independent variables based on the energy function, using the gradient descent method and a fixed step size along the negative gradient direction.

[0080] The frequency determination module is used to output the corresponding angular frequency as the frequency of the oscillation signal under test when the gradient of the energy function is zero.

[0081] The present invention also provides a computer-readable storage medium storing a program that, when executed by a processor, implements the above-described method.

[0082] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; only preferred embodiments of the present invention are illustrated. The descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. As long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0083] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A method for detecting a subsynchronous oscillation based on a tunable notch filter, characterized by: The method comprises the following steps: Step S1: Cascading a direct current digital filter and a tunable digital notch filter to obtain a single-frequency oscillation signal detector; Step S2: Taking the notch angular frequency of the tunable notch filter as an independent variable, and constructing an energy function based on the output of the single-frequency oscillation signal detector and the to-be-detected oscillation signal; Step S3: Using a gradient descent method to iteratively update the independent variable in the negative gradient direction of the energy function at a fixed step length, and searching for the minimum value of the energy function: when the gradient of the energy function with respect to the independent variable is zero, the corresponding angular frequency is determined as the angular frequency of the to-be-detected oscillation signal; The energy function The expression for the energy function is ； In the formula, is the independent variable after the system is discretized; represents the accumulated delay number of a single calculation; represents the current time; represents a single-frequency oscillation signal detector; represents the value of the single-frequency oscillation signal detector at the s beat; is the to-be-tested oscillation signal input into the single-frequency oscillation signal detector.

2. The method according to claim 1, wherein the method is based on a tunable notch filter. Expression of the tunable digital notch filter is: ； wherein is the sampling interval; is the system discretization argument; is the notch corner frequency.

3. The method of sub-synchronous oscillation detection based on a tunable notch filter according to claim 2, characterized in that: Expression of the direct current digital filter is: 。 4. The method of sub-synchronous oscillation detection based on tunable notch filter according to claim 1, characterized in that: accumulating a number of delays for the single calculation The following conditions are set: ； In the formula, represents a sampling interval; represents a notch frequency.

5. The method of sub-synchronous oscillation detection based on tunable notch filter according to claim 1, characterized in that: When the to-be-detected oscillation signal is a multi-frequency oscillation signal, first, segment filtering preprocessing is performed, and then the frequency of the oscillation component in different frequency bands is detected through step S3.

6. A tunable trap based sub-synchronous oscillation detection system, implemented based on the method of any one of claims 1 to 5, characterized in that: The method comprises a direct current digital filtering module, a tunable digital notch filtering module, a cascaded detector module, an energy function calculation module, an iterative calculation module, and a frequency determination module; The direct current digital filtering module is configured to filter the input to-be-detected signal to eliminate direct current bias; The tunable digital notch filtering module is configured to perform notch processing on the to-be-detected signal, so that when the notch frequency is consistent with the frequency of the to-be-detected single-frequency sinusoidal signal, the output is zero; The cascaded detector module is formed by connecting the direct current digital filtering module and the tunable digital notch filtering module in series, and is configured to output a single-frequency oscillation signal detection result; The energy function calculation module is configured to take the notch angular frequency of the tunable digital notch filtering module as an independent variable, and construct an energy function based on the output of the cascaded detector module; The iterative calculation module is configured to iteratively update the independent variable in the negative gradient direction based on the energy function using a gradient descent method at a fixed step length; The frequency determination module is configured to output the corresponding angular frequency as the frequency of the to-be-detected oscillation signal when the gradient of the energy function is zero.

7. A computer-readable storage medium, characterized in that: The program is stored in the memory and is executed by the processor to implement the method of any one of claims 1 to 5.

Images