Data-based network construction type power supply inertia estimation method

Abstract

The invention relates to the technical field of network construction type converter-based power supply inertia estimation based on small disturbance response, in particular to a data-based network construction type power supply inertia estimation method. Comprising the following steps: obtaining an inertia response of a network-forming CBG; obtaining a network construction type CBG inertia expression based on electromechanical characteristics; and extracting the electromechanical oscillation component. The inertia response model under the electromechanical time scale is constructed on the basis of analyzing the time scale of each control link in the detailed model of the network construction type CBG. By analyzing the electromechanical oscillation response characteristics of the net-forming type CBG under small disturbance, the analytical expression of the inertia of the net-forming type CBG is deduced, and the situation that the inertia is evaluated to be negative can be effectively avoided. A dominant oscillation mode is extracted through an oscillation mode decomposition technology, and the application of the estimation method in a large-scale multi-machine system is improved.

Description

Technical Field

[0001] This invention relates to the field of grid-based converter power supply inertia estimation technology based on small disturbance response, and particularly to a data-based grid-based power supply inertia estimation method. Background Technology

[0002] With a large amount of new energy sources being connected to the power grid via power electronic converters, the inertia level of the power system is continuously decreasing, and the uneven distribution of inertia is becoming more pronounced. Inertia is an inherent property of the power system and one of the important parameters for measuring the stable operation of the system, significantly affecting system stability. Therefore, accurately understanding the inertia characteristics of the power system is of great significance.

[0003] Early inertia estimation methods primarily targeted power systems dominated by synchronous generators, using the dynamic model of the synchronous machine to analyze the system inertia as the ratio of the power deficit to the rate of frequency change (RoCoF) at the moment of disturbance after a transient accident. With the increasing maturity of wide-area measurement technology, electrical variables such as generator power and speed can be more easily obtained, allowing for the assessment of system inertia by identifying the system model. However, this method requires the injection of additional probe signals, increasing implementation complexity. The N4SID algorithm, ARMAX algorithm, and error moving average can be used to identify the system model, extracting inertia from the model's step response. Inertia estimation can be performed by training switching Markov Gaussian models and artificial neural network models using transient historical data. In summary, existing technologies require a method for assessing the equivalent inertia of regional power grids to improve the accuracy of the assessment results.

[0004] With the development of converters, grid-type converters will be widely adopted. Since the advantage of grid-type converters is that they can fundamentally simulate the inertia of synchronous generators, the inertia analysis of grid-type power generation will be conducted as part of the power system inertia estimation. Summary of the Invention

[0005] The technical problem this invention aims to solve is that, with the development of converters, grid-type converters will be widely adopted. Since the advantage of grid-type converters is their ability to fundamentally simulate the inertia of synchronous generators, the inertia analysis of grid-type power generation will be conducted as part of the power system inertia estimation.

[0006] The inertia and damping coefficient of a regional power grid are inherent equivalent inertia of the power system and participate in the construction of the power system's state-space model. Through characteristic analysis of the state-space model, the correlation between equivalent inertia and inherent characteristic information of electromechanical dynamic response can be established. From the perspective of inertia assessment, how to establish a state-space model through a data-driven approach is key to establishing the relationship between equivalent inertia and characteristic information.

[0007] A data-based method for estimating the inertia of a networked power supply, the method comprising the following steps:

[0008] Step 1: Obtain the inertia response of the meshed CBG.

[0009] Step 2: Obtain the inertia expression of the networked CBG based on electromechanical characteristics.

[0010] Step 3: Extract the electromechanical oscillation component.

[0011] The basic principle of the inertial response of the mesh-type CBG in step 1 is as follows:

[0012] Since grid-connected CBGs can sense disturbances to the power grid, they participate in the grid's electromechanical response process after being connected to the grid, thus causing a shift in state variables. Therefore:

[0013]

[0014] By reversing equation (1), the moment of inertia can be expressed as:

[0015]

[0016] After obtaining ΔPe and Δω, inertia can be directly evaluated using equation (1). However, one problem that needs to be explored in this evaluation is whether there is a numerator or denominator that is zero. Therefore, it is necessary to determine the time-domain characteristics of the responses ΔPe and Δω. The existence of virtual impedance control means that there is a constant phase angle difference between the potential within the grid-type CBG and its outlet voltage, resulting in power loss. Referring to the linearization process of a synchronous generator, the variable ΔPe in equation (1) can be linearly expressed as:

[0017] ΔP e =k s Δθ (3)

[0018] Where k s These are the potential and terminal voltage power transmission coefficients within the grid-type CBG; by substituting equation (3) into equation (1), we can obtain the second-order electromechanical model of the grid-type CBG, namely:

[0019]

[0020] The form of the response can be determined by analyzing the general solution of the differential equation shown in equation (4). According to the eigenvalue analysis theory, the characteristic equation corresponding to equation (4) is solved as a pair of conjugate complex roots. The real part of the eigenvalue corresponds to the damping ratio of the electromechanical oscillation response process, and the imaginary part corresponds to the oscillation frequency.

[0021]

[0022] As can be seen from the formula, the response of the network-type CBG after being disturbed exhibits an oscillating process. From a general perspective, the oscillation process can be described by a decaying signal, that is, Δω follows the following form:

[0023]

[0024] Among them, A ω The frequency oscillation amplitude, The initial phase is the frequency. Although there is a differential relationship between Δθ and Δω, due to the characteristics of the derivatives of trigonometric and exponential functions, Δθ still follows the form of a decaying signal. And Δθ and ΔP... e There is a linear relationship between them, while Δθ and ΔP e There exists a linear relationship between them. Therefore, we have:

[0025]

[0026] It can be seen that Δω and ΔP e Both equations alternate around the zero axis, meaning that the value of equation (2) is likely to be 0 in either the numerator or denominator, thus the calculated inertia will not satisfy practical significance. From equations (6) and (7), it can be seen that although Δω or ΔP... e It varies around the 0 axis, but it has a damping term, which means that the closed region it forms is not zero within one oscillation period, and this closed region corresponds to the integration operation. Therefore, the expression for the inertia can be obtained by integrating both sides of equation (3), that is:

[0027]

[0028] In the formula: t1 represents the upper limit of the integration interval; the upper limit of the integration interval can be selected at any time within the oscillation window, but in order to fully explore the inertia information in the data and improve the evaluation accuracy and average evaluation error, the upper limit of integration is selected at the time when the time window ends.

[0029] The basic principle of step 2, extraction of electromechanical oscillation components, is as follows:

[0030] A method for extracting characteristic parameters of electromechanical oscillations based on adaptive local iterative filters. Its basic principle is described by the following filtering process:

[0031]

[0032] Where z(t) is the multimode coupled signal; χ is the wave operator, χ(z(t))=z(t)-κ(z(t)), where κ is the sliding operator. According to (9), the rotational speed and power IMF oscillation components used for inertia assessment can be obtained through the following screening process:

[0033] Step 1: Calculate the sliding operator κ(z(t)) of the multimode coupled signal z;

[0034] Step 2: Remove the sliding operator from the original signal z(t) to obtain the wave operator χ(z(t));

[0035] Step 3: Determine whether the fluctuation operator calculated for χ(z(t)) meets the conditions. If the conditions are not met, repeat the calculation of the fluctuation operator using steps 1 and 2. If the conditions are met, the fluctuation operator is the IMF component.

[0036] Step 4: Subtract the original signal z(t) from the extracted fluctuation operator χ(z(t)) that satisfies the IMF component, calculate the residual component rt, and repeat steps 1-4 as the new initial signal until the residual component is a significant trend component; after a finite number of iterations, the IMF components of different frequencies contained in the original signal are obtained; finally, the original signal is decomposed into the superposition of multiple IMF components with modal singularities and trend quantities.

[0037]

[0038] Beneficial effects

[0039] Through the above design scheme, this invention can bring the following beneficial effects: This paper proposes an inertia estimation method for networked CBG based on small disturbance response. Based on the time-scale analysis of each control element in the detailed model of the networked CBG, an inertia response model under the electromechanical time scale is constructed. By analyzing the electromechanical oscillation response characteristics of the networked CBG under small disturbances, an analytical expression for its inertia is derived, effectively avoiding the risk of complex inertia estimation. The dominant oscillation mode is extracted using oscillation mode decomposition technology, improving the application of the estimation method in large-scale multi-machine systems.

[0040] This method is based on the relationship between inertia and electromechanical oscillation characteristics, estimating inertia by extracting and averaging the oscillation components. This project focuses on the inertia analysis of grid-type power sources, simplifying the time-scale analysis of detailed models while preserving the electromechanical time-scale relationship of inertia parameters, and establishing a dynamic model. Based on the dynamic model, inertia analysis is derived, and considering the response characteristics of grid-type power sources, an inertia estimation method based on data averaging is proposed. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings involved in the embodiments or the prior art are briefly described below. Obviously, these drawings illustrate several embodiments of the present invention, and those skilled in the art can derive other possible drawings based on these drawings without creative effort. The purpose of the drawings is limited to illustrating specific embodiments and does not limit the scope of the present invention.

[0042] Figure 1 Control block diagram of a grid-type converter;

[0043] Figure 2 Active power control block diagram;

[0044] Figure 3 Reactive power control block diagram;

[0045] Figure 4 Block diagram of voltage and current dual closed-loop control;

[0046] Figure 5 Virtual impedance control block diagram;

[0047] Figure 6 Schematic diagram of an IEEE 4-machine system;

[0048] Figure 7 (a) The spectrum of the velocity signal;

[0049] Figure 7 (b) Spectrum of power signal. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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.

[0051] A data-based method for estimating the inertia of a networked power supply, the method comprising the following steps:

[0052] Step 1: Obtain the inertia response of the meshed CBG.

[0053] Step 2: Obtain the inertia expression of the networked CBG based on electromechanical characteristics.

[0054] Step 3: Extract the electromechanical oscillation component.

[0055] The basic principle of the inertial response of the mesh-type CBG in step 1 is as follows:

[0056] Since grid-connected CBGs can sense disturbances to the power grid, they participate in the grid's electromechanical response process after being connected to the grid, thus causing a shift in state variables. Therefore:

[0057]

[0058] By reversing equation (1), the moment of inertia can be expressed as:

[0059]

[0060] After obtaining ΔPe and Δω, inertia can be directly evaluated using equation (1). However, one problem that needs to be explored in this evaluation is whether there is a numerator or denominator that is zero. Therefore, it is necessary to determine the time-domain characteristics of the responses ΔPe and Δω. The existence of virtual impedance control means that there is a constant phase angle difference between the potential within the grid-type CBG and its outlet voltage, resulting in power loss. Referring to the linearization process of a synchronous generator, the variable ΔPe in equation (1) can be linearly expressed as:

[0061] ΔP e =k s Δθ (3)

[0062] Where k s These are the potential and terminal voltage power transmission coefficients within the grid-type CBG; by substituting equation (3) into equation (1), we can obtain the second-order electromechanical model of the grid-type CBG, namely:

[0063]

[0064] The form of the response can be determined by analyzing the general solution of the differential equation shown in equation (4). According to the eigenvalue analysis theory, the characteristic equation corresponding to equation (4) is solved as a pair of conjugate complex roots. The real part of the eigenvalue corresponds to the damping ratio of the electromechanical oscillation response process, and the imaginary part corresponds to the oscillation frequency.

[0065]

[0066] As can be seen from the formula, the response of the network-type CBG after being disturbed exhibits an oscillating process. From a general perspective, the oscillation process can be described by a decaying signal, that is, Δω follows the following form:

[0067]

[0068] Among them, A ω The frequency oscillation amplitude, The initial phase is the frequency. Although there is a differential relationship between Δθ and Δω, due to the characteristics of the derivatives of trigonometric and exponential functions, Δθ still follows the form of a decaying signal. And Δθ and ΔP... e There is a linear relationship between them, while Δθ and ΔP e There exists a linear relationship between them. Therefore, we have:

[0069]

[0070] It can be seen that Δω and ΔP eBoth equations alternate around the zero axis, meaning that the value of equation (2) is likely to be 0 in either the numerator or denominator, thus the calculated inertia will not satisfy practical significance. From equations (6) and (7), it can be seen that although Δω or ΔP... e It varies around the 0 axis, but it has a damping term, which means that the closed region it forms is not zero within one oscillation period, and this closed region corresponds to the integration operation. Therefore, the expression for the inertia can be obtained by integrating both sides of equation (3), that is:

[0071]

[0072] In the formula: t1 represents the upper limit of the integration interval; the upper limit of the integration interval can be selected at any time within the oscillation window, but in order to fully explore the inertia information in the data and improve the evaluation accuracy and average evaluation error, the upper limit of integration is selected at the time when the time window ends.

[0073] The basic principle of step 2, extraction of electromechanical oscillation components, is as follows:

[0074] A method for extracting characteristic parameters of electromechanical oscillations based on adaptive local iterative filters. Its basic principle is described by the following filtering process:

[0075]

[0076] Where z(t) is the multimode coupled signal; χ is the wave operator, χ(z(t))=z(t)-κ(z(t)), where κ is the sliding operator. According to (9), the rotational speed and power IMF oscillation components used for inertia assessment can be obtained through the following screening process:

[0077] Step 1: Calculate the sliding operator κ(z(t)) of the multimode coupled signal z;

[0078] Step 2: Remove the sliding operator from the original signal z(t) to obtain the wave operator χ(z(t));

[0079] Step 3: Determine whether the fluctuation operator calculated for χ(z(t)) meets the conditions. If the conditions are not met, repeat the calculation of the fluctuation operator using steps 1 and 2. If the conditions are met, the fluctuation operator is the IMF component.

[0080] Step 4: Subtract the original signal z(t) from the extracted fluctuation operator χ(z(t)) that satisfies the IMF component, calculate the residual component rt, and repeat steps 1-4 as the new initial signal until the residual component is a significant trend component; after a finite number of iterations, the IMF components of different frequencies contained in the original signal are obtained; finally, the original signal is decomposed into the superposition of multiple IMF components with modal singularities and trend quantities.

[0081]

[0082] This invention extracts the dominant oscillation modes through oscillation mode decomposition technology, improving the application of estimation methods in large-scale multi-machine systems. The effectiveness of the proposed method is verified through numerical examples of an IEEE 4-machine power system. Specific details are as follows:

[0083] Taking the IEEE 4-machine system as an example, the network structure of the IEEE 4-machine 11-node system is as follows: Figure 6 The diagram shows a typical two-zone interconnected system. Based on the original system, the synchronous generator corresponding to G4 is replaced with a CBG, and the inertia of the CBG is set to 15. A three-phase short-circuit fault of 0.05s is applied to bus 10. Figure 7 The power deviation from CBG, ΔPe in Equation (10), and the speed signal deviation, Δω in Equation (10), are shown. The speed signal exhibits a clear trend component. Spectral analysis of the speed and power signals shows that both signals contain two oscillation modes, with the mode at 0.5098 Hz being more dominant, in addition to the coupling of the oscillation modes.

[0084] After applying the ALIFD algorithm, components with an oscillation frequency of 0.5098 Hz were extracted from the speed signal and power signal, respectively. The extracted oscillation components were substituted into equation (10) for evaluation, and the results are shown in Table 1. Compared with the actual value, the estimated value still deviates from the actual value, but the deviation is small, only 1.9. From the perspective of error, the relative error is 12.67%, which is within the acceptable range.

[0085] Table 1. Inertia Assessment Results

[0086]

[0087] Table 2. Inertia Estimation Results

[0088]

[0089] Those skilled in the art should understand that the above embodiments are merely illustrative of the content of this disclosure and do not limit its scope. The system capacity, voltage, line parameters, etc., shown may vary depending on the specific circumstances of the power electronic grid-connected generator set and its grid connection. Based on this disclosure, those skilled in the art can make other changes or adjustments, and these changes still fall within the scope of this disclosure.

Claims

1. A data-based method for estimating the inertia of a networked power supply, characterized in that, Includes the following steps: Step 1: Obtain the inertial response of the meshed CBG; Step 2: Obtain the inertia expression for the networked CBG based on electromechanical characteristics; Step 3: Extract the electromechanical oscillation component.

2. The data-based network-type power supply inertia estimation method according to claim 1, characterized in that, Step 1: Obtain the inertial response of the meshed CBG, which specifically includes: A grid-connected CBG can sense disturbances to the power grid. When connected to the grid, it participates in the grid's electromechanical response process, causing a shift in state variables. Therefore: By reversing equation (1), the moment of inertia can be expressed as: Obtain ΔP e and After that, the inertia can be directly evaluated using the formula.

3. The data-based network-type power supply inertia estimation method according to claim 2, characterized in that, It also includes: determining ΔP e and The time-domain characteristics specifically include the following steps: The existence of virtual impedance control implies a constant phase angle difference between the potential within the grid-type CBG and its output terminal voltage, resulting in power loss; referring to the linearization process of a synchronous generator, the variable ΔPe in the equation can be linearly expressed as: ΔP e =k s Dth (3) Where k s These are the potential and terminal voltage power transmission coefficients within the grid-type CBG; by substituting equation (3) into equation (1), we can obtain the second-order electromechanical model of the grid-type CBG, namely: The form of the response can be determined by analyzing the general solution of the differential equation shown in equation (4).

4. The data-based network-type power supply inertia estimation method according to claim 3, characterized in that, Step 2: Obtain the inertia expression for the networked CBG based on electromechanical characteristics, specifically... Includes the following steps: According to the characteristic analysis theory, the characteristic equation corresponding to equation (4) is solved as a pair of conjugate complex roots. The real part of the eigenvalue corresponds to the damping ratio of the electromechanical oscillation response process, and the imaginary part corresponds to the oscillation frequency. As can be seen from the formula, the response of the network-type CBG after being disturbed exhibits an oscillating process. From a general perspective, the oscillation process can be described by a decaying signal, that is, Δω follows the following form: Among them, A ω The frequency oscillation amplitude, The initial phase is the frequency; Δθ follows the form of a decaying signal, and Δθ and ΔP are related. e There is a linear relationship between them, therefore ΔP e It also conforms to the attenuation signal form: Δω or ΔP e The closed region formed by the change around the 0 axis is not zero within one oscillation period, and this closed region corresponds to the integration operation.

5. The data-based network-type power supply inertia estimation method according to claim 4, characterized in that, Also includes: By integrating both sides of equation (3) simultaneously, we obtain the expression for the inertia, namely: In the formula: t1 represents the upper limit of the integration interval; The upper limit of the integration interval can be chosen at any time within the oscillation window, but in order to fully explore the inertia information in the data and improve the evaluation accuracy and average evaluation error, the upper limit of integration is chosen at the end of the time window.

6. The data-based network-type power supply inertia estimation method according to claim 1, characterized in that, Step 3: Extraction of electromechanical oscillation components, which specifically includes: The method for extracting electromechanical oscillation characteristic parameters based on adaptive local iterative filters is described by the following filtering process: Where z(t) is a multimode coupled signal; χ is a wave operator, χ(z(t))=z(t)-κ(z(t)), where κ is a sliding operator.

7. The data-based network-type power supply inertia estimation method according to claim 6, characterized in that, Also includes: According to (9), the rotational speed and power IMF oscillation components used for inertia assessment can be obtained through the following screening process: Step 1: Calculate the sliding operator κ(z(t)) of the multimode coupled signal z; Step 2: Remove the sliding operator from the original signal z(t) to obtain the wave operator χ(z(t)); Step 3: Determine whether the fluctuation operator calculated for χ(z(t)) meets the conditions. If the conditions are not met, repeat the calculation of the fluctuation operator using steps 1 and 2. If the conditions are met, the fluctuation operator is the IMF component. Step 4: Subtract the original signal z(t) from the extracted fluctuation operator χ(z(t)) that satisfies the IMF component, calculate the residual component rt, and repeat steps 1-4 as the new initial signal until the residual component is a significant trend component. After a finite number of iterations, the IMF components of different frequencies contained in the original signal were obtained; finally, the original signal was decomposed into a superposition of multiple IMF components with modal singularities and trend quantities.

Images