Numerical oscillation suppression method and system for electromagnetic transient simulation based on eigenvalue analysis

Abstract

The present invention relates to the field of power systems and discloses a method for suppressing numerical oscillations in electromagnetic transient simulation based on eigenvalue analysis, comprising the following steps: using an electromagnetic transient model to discretize characteristic formulas of components with memory in a circuit into linear conductance and historical current sources; during normal simulation calculations, using an implicit trapezoidal method to calculate state variables and eigenvalues ​​of the electromagnetic transient model; after a circuit breaker is operated, calculating the electromagnetic transient eigenvalues ​​of the circuit under the back-step Euler method discretization, and calculating, based on the eigenvalues, how many steps of half-step back-step Euler method calculations are required to completely suppress numerical oscillations; reducing the simulation step size to 1 / 2 of the original simulation step size, performing k steps of half-step back-step Euler method calculations; after completing the k steps of half-step back-step Euler method calculations, switching back to the implicit trapezoidal method to continue the calculations, thereby completing the suppression of numerical oscillations in electromagnetic transient simulation. The present invention can more effectively suppress numerical oscillations after circuit breaker operation.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and in particular to an electromagnetic transient simulation numerical oscillation suppression method and system based on eigenvalue analysis. Background Art

[0002] Electromagnetic transient simulation is a crucial tool for studying the dynamic characteristics of power systems. Compared to electromechanical transient simulation, it can more accurately describe the dynamic characteristics of power electronic systems. Currently, mature electromagnetic transient simulation analysis software includes EMTP, PSCAD / EMTDC developed by the Manitoba DC Research Center in Canada, NETOMAC developed by Siemens in Germany, and CLOUDPSS developed by the Institute of Digital Twins of Energy and Power Systems at the Tsinghua Sichuan Energy Internet Research Institute.

[0003] The implicit trapezoidal integration method is widely used in electromagnetic transient simulations. It is a very effective, highly stable, and accurate numerical integration method. However, when a circuit breaker is operated in the network, this method can cause abnormal fluctuations in non-state variables, a phenomenon known as numerical oscillation. This abnormal oscillation phenomenon, if not eliminated, will affect the accuracy of subsequent simulations. The root cause of numerical oscillation is that when some form of disturbance occurs in the network (such as a circuit breaker operation), the non-state variables in the network undergo a sudden change. However, since the non-state variables after the sudden change are difficult to obtain, the pre-change variables are often used to calculate the equivalent current source after the sudden change. This calculated equivalent current source is obviously incorrect, which is reflected in the calculation results as abnormal fluctuations in the non-state variables. Generally, numerical oscillation is easily generated when a circuit breaker is operated, and research has shown that short-circuit faults can also cause numerical oscillation in certain circumstances.

[0004] Currently, the most widely used method for suppressing numerical oscillations is the critical damping adjustment (CDA) method. This involves switching the numerical integration method to the backward Euler method at the instant of circuit breaker operation, performing a two-step simulation. The backward Euler method avoids numerical oscillations because its differential equations do not contain non-state variables at that moment, thus preventing sudden changes in non-state variables from creating undesirable equivalent injection sources. However, the backward Euler method is less accurate than the trapezoidal method and is not suitable for use throughout the entire simulation. Typically, the backward Euler method is used for a few steps after a disturbance such as circuit breaker operation, followed by a switch to the trapezoidal method.

[0005] While the CDA method can suppress numerical oscillations to a certain extent, it also has significant drawbacks. Suppressing numerical oscillations for different circuit networks requires performing Backward Euler calculations with different numbers of steps. The CDA method performs a two-step Backward Euler calculation on all circuit networks. This reduces simulation accuracy for networks requiring fewer than two Backward Euler steps, and fails to effectively suppress numerical oscillations for networks requiring more than two steps. Summary of the Invention

[0006] The present invention aims to overcome the shortcomings of the prior art and provide a method for suppressing numerical oscillations in electromagnetic transient simulation based on eigenvalue analysis, comprising the following steps:

[0007] Step 1: The electromagnetic transient model discretizes the characteristic formulas of the components with memory in the circuit into linear conductance and historical current sources. During normal simulation calculations, the implicit trapezoidal method is used to calculate the state variables and eigenvalues ​​of the electromagnetic transient model. The circuit state variables and the eigenvalues ​​of the electromagnetic transient model at the previous moment determine the value of the historical current source.

[0008] Step 2: After the circuit breaker is operated, the electromagnetic transient characteristic value of the circuit is calculated under the backward Euler method discretization, and the half-step backward Euler method calculation of k steps that completely suppresses the numerical oscillation phenomenon is obtained based on the electromagnetic transient characteristic value; the simulation step size is reduced to the original simulation step size. , calculate the historical current source of each component according to the backward Euler method, where the linear conductance value remains unchanged, and perform k steps of half-step backward Euler method calculation;

[0009] Step 3: After performing k steps of half-step backward Euler method calculation, switch back to the implicit trapezoidal method to continue the calculation and complete the numerical oscillation suppression of the magnetic transient simulation.

[0010] Furthermore, the circuit contains memory elements, and the memory elements include inductors and capacitors.

[0011] Furthermore, during normal simulation calculations, the implicit trapezoidal method is used to calculate the eigenvalues ​​of the electromagnetic transient model. After the circuit breaker is operated, the backward Euler method is used to calculate the eigenvalues ​​of the electromagnetic transient model. The circuit state variables at the previous moment and the eigenvalues ​​of the electromagnetic transient model determine the value of the historical current source, including:

[0012] In RL circuit, for the resistor R have

[0013]

[0014] For inductors L have

[0015]

[0016] For inductance L Trapezoidal discretization,

[0017]

[0018]

[0019] get

[0020]

[0021] in

[0022] Then the characteristic equation of the inductor L after discretization using the implicit trapezoidal method is:

[0023]

[0024] The calculation formula of the historical current source is:

[0025]

[0026] Then there is

[0027]

[0028] get

[0029]

[0030] but

[0031]

[0032] get

[0033]

[0034] Then the implicit trapezoidal electromagnetic transient characteristic value of the RL circuit is:

[0035]

[0036] Eigenvalues ​​under the backward Euler discretization

[0037] RL circuit

[0038] For resistors R have

[0039]

[0040] For inductors L have

[0041]

[0042] Discretization using the Backward Euler method

[0043]

[0044] Arranged

[0045]

[0046] in ;

[0047] Then the characteristic equation of the inductor L after discretization using the backward Euler method is:

[0048]

[0049] in

[0050]

[0051] have

[0052]

[0053] Arranged

[0054]

[0055] but

[0056]

[0057] get

[0058]

[0059] That is, the backward Euler method electromagnetic transient eigenvalue of the RL circuit is:

[0060] .

[0061] Furthermore, after the circuit breaker is operated, the electromagnetic transient characteristic value of the circuit is calculated under the backward Euler method discretization, and based on the electromagnetic transient characteristic value, a half-step backward Euler method calculation of k steps that completely suppresses the numerical oscillation phenomenon is obtained; the simulation step size is reduced to the original simulation step size. , calculate the historical current source of each component according to the backward Euler method, where the linear conductance value remains unchanged, and perform k steps of half-step backward Euler method calculation, including:

[0062] If used After the half-step backward Euler method, the numerical oscillation is effectively suppressed. The historical current source value after the first half-step backward Euler method is:

[0063]

[0064] The historical current source value after the second half-step backward Euler method is:

[0065]

[0066] After performing the k-step backward Euler method calculation, the historical current source value is:

[0067]

[0068] when Decay to of times, ,Right now

[0069]

[0070] but

[0071]

[0072] get

[0073]

[0074] but

[0075]

[0076] The minimum number of steps for:

[0077]

[0078] The oscillating voltage after k steps of half-step backward Euler method calculation is:

[0079]

[0080]

[0081] Then proceed The numerical oscillation voltage after the half-step backward Euler method is:

[0082] .

[0083] An electromagnetic transient simulation numerical oscillation suppression system based on eigenvalue analysis applies the electromagnetic transient simulation numerical oscillation suppression method based on eigenvalue analysis, and includes an electromagnetic transient simulation module, a data processing module, and a data storage module; the electromagnetic transient simulation module and the data storage module are respectively connected to the data processing module.

[0084] The beneficial effects of this invention are as follows: Based on eigenvalue analysis, this patent proposes a novel method that is more effective than traditional methods for suppressing numerical oscillations in electromagnetic transient simulations and can lessen the impact on computational accuracy. Compared with traditional methods for suppressing numerical oscillations in electromagnetic transient simulations, this method can more effectively suppress the numerical oscillation phenomenon after circuit breaker operation and can lessen the impact on computational accuracy, thus improving and refining the electromagnetic transient simulation algorithm and enhancing the practicality and universality of electromagnetic transient simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 The figure is a flow chart of the numerical oscillation suppression method for electromagnetic transient simulation based on eigenvalue analysis;

[0086] Figure 2 is the schematic diagram of RL circuit;

[0087] Figure 3 This is the RL circuit diagram after the trapezoidal method is discretized;

[0088] Figure 4 This is the discrete circuit diagram after R is disconnected;

[0089] Figure 5 Schematic diagram of the voltage across the inductor calculated using the implicit trapezoidal method;

[0090] Figure 6 Schematic diagram of the inductor historical current source value calculated using the implicit trapezoidal method;

[0091] Figure 7 Schematic diagram of the voltage across the inductor calculated using the traditional numerical oscillation suppression method;

[0092] Figure 8 Schematic diagram of the inductor historical current source value calculated using the traditional numerical oscillation suppression method;

[0093] Figure 9 Schematic diagram of the voltage across the inductor calculated using the numerical oscillation suppression method for electromagnetic transient simulation based on eigenvalue analysis;

[0094] Figure 10 The inductor historical current source value calculated by the numerical oscillation suppression method for electromagnetic transient simulation based on eigenvalue analysis. DETAILED DESCRIPTION

[0095] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0096] The features and performance of the present invention are further described in detail below with reference to the embodiments.

[0097] like Figure 1As shown in FIG, the numerical oscillation suppression method for electromagnetic transient simulation based on eigenvalue analysis includes the following steps:

[0098] Step 1: The electromagnetic transient model discretizes the characteristic formulas of the components with memory in the circuit into linear conductance and historical current sources. During normal simulation calculations, the implicit trapezoidal method is used to calculate the state variables and eigenvalues ​​of the electromagnetic transient model. The circuit state variables and the eigenvalues ​​of the electromagnetic transient model at the previous moment determine the value of the historical current source.

[0099] Step 2: After the circuit breaker is operated, the electromagnetic transient characteristic value of the circuit is calculated under the backward Euler method discretization, and the half-step backward Euler method calculation of k steps that completely suppresses the numerical oscillation phenomenon is obtained based on the electromagnetic transient characteristic value; the simulation step size is reduced to the original simulation step size. , calculate the historical current source of each component according to the backward Euler method, where the linear conductance value remains unchanged, and perform k steps of half-step backward Euler method calculation;

[0100] Step 3: After performing k steps of half-step backward Euler method calculation, switch back to the implicit trapezoidal method to continue the calculation and complete the numerical oscillation suppression of the magnetic transient simulation.

[0101] The circuit contains memory elements, which include inductors and capacitors.

[0102] During normal simulation, the implicit trapezoidal method is used to calculate the eigenvalues ​​of the electromagnetic transient model. After the circuit breaker is operated, the backward Euler method is used to calculate the eigenvalues ​​of the electromagnetic transient model. The circuit state variables at the previous moment and the eigenvalues ​​of the electromagnetic transient model determine the value of the historical current source, including:

[0103] In RL circuit, for the resistor R have

[0104]

[0105] For inductors L have

[0106]

[0107] For inductance L Trapezoidal discretization,

[0108]

[0109]

[0110] get

[0111]

[0112] in

[0113] Then the characteristic equation of the inductor L after discretization using the implicit trapezoidal method is:

[0114]

[0115] The calculation formula of the historical current source is:

[0116]

[0117] Then there is

[0118]

[0119] get

[0120]

[0121] but

[0122]

[0123] get

[0124]

[0125] Then the implicit trapezoidal electromagnetic transient characteristic value of the RL circuit is:

[0126]

[0127] Eigenvalues ​​under the backward Euler discretization

[0128] RL circuit

[0129] For resistors R have

[0130]

[0131] For inductors L have

[0132]

[0133] Discretization using the Backward Euler method

[0134]

[0135] Arranged

[0136]

[0137] in ;

[0138] Then the characteristic equation of the inductor L after discretization using the backward Euler method is:

[0139]

[0140] in

[0141]

[0142] have

[0143]

[0144] Arranged

[0145]

[0146] but

[0147]

[0148] get

[0149]

[0150] That is, the backward Euler method electromagnetic transient eigenvalue of the RL circuit is:

[0151] .

[0152] After the circuit breaker is operated, the electromagnetic transient characteristic value of the circuit is calculated under the backward Euler method discretization, and the half-step backward Euler method calculation of k steps that completely suppresses the numerical oscillation phenomenon is obtained based on the electromagnetic transient characteristic value; the simulation step size is reduced to the original simulation step size , calculate the historical current source of each component according to the backward Euler method, where the linear conductance value remains unchanged, and perform k steps of half-step backward Euler method calculation, including:

[0153] If used After the half-step backward Euler method, the numerical oscillation is effectively suppressed. The historical current source value after the first half-step backward Euler method is:

[0154]

[0155] The historical current source value after the second half-step backward Euler method is:

[0156]

[0157] After performing the k-step backward Euler method calculation, the historical current source value is:

[0158]

[0159] when Decay to of times, ,Right now

[0160]

[0161] but

[0162]

[0163] get

[0164]

[0165] but

[0166]

[0167] The minimum number of steps for:

[0168]

[0169] The oscillating voltage after k steps of half-step backward Euler method calculation is:

[0170]

[0171]

[0172] Then proceed The numerical oscillation voltage after the half-step backward Euler method is:

[0173] .

[0174] An electromagnetic transient simulation numerical oscillation suppression system based on eigenvalue analysis applies the electromagnetic transient simulation numerical oscillation suppression method based on eigenvalue analysis, and includes an electromagnetic transient simulation module, a data processing module, and a data storage module; the electromagnetic transient simulation module and the data storage module are respectively connected to the data processing module.

[0175] Specifically, to determine the exact number of backward Euler steps required for different circuit networks after a circuit breaker operation, we considered using the eigenvalues ​​of the electromagnetic transient model. The electromagnetic transient model discretizes the characteristic formulas of memory components, such as inductors and capacitors, into a linear conductance and a historical current source. The value of the historical current source is determined by the circuit state variables at the previous moment and the eigenvalues ​​of the electromagnetic transient model. The eigenvalues ​​of the electromagnetic transient model reflect the changes in the historical current source and the system characteristics during the simulation.

[0176] Eigenvalues ​​under the backward Euler discretization

[0177] Figure 2The circuit shown in derives and analyzes the historical current source eigenvalues ​​obtained by discretizing the RL circuit using the backward Euler method.

[0178] For resistors R have

[0179] (1-1)

[0180] For inductors L have

[0181] (1-2)

[0182] Because the voltage value across the resistor is only related to the current value at this moment, and has nothing to do with the current value at the previous moment, and the two are linearly related, the historical current source of the resistor is always 0, that is,

[0183] The inductance characteristic formula of Equation (1-2) is discretized using the backward Euler method.

[0184] (1-3)

[0185] Arranged

[0186] (1-4)

[0187] in

[0188] Then the characteristic equation of the inductor L after discretization using the backward Euler method is:

[0189] (1-5)

[0190] in

[0191] (1-6)

[0192] like Figure 3 Discrete RL circuit diagram, using Kirchhoff's law for node 1, we have

[0193] (1-7)

[0194] Arranged

[0195] (1-8)

[0196] Since the right end of the inductor is grounded, the voltage across the inductor is equal to the voltage at node 1.

[0197] (1-9)

[0198] Substituting (1-9) into (1-6) we get

[0199] (1-10)

[0200] That is, the backward Euler method electromagnetic transient characteristic value of the RL circuit is

[0201] (1-11)

[0202] Eigenvalues ​​under implicit trapezoidal discretization

[0203] by Figure 2 The circuit shown in the derivation and analysis uses the implicit ladder method to discretize the RL circuit to obtain the historical current source eigenvalues.

[0204] For resistors R have

[0205] (1-12)

[0206] For inductors L have

[0207] (1-13)

[0208] and Figure 2 Because the voltage value across the resistor is only related to the current value at this moment, and has nothing to do with the current value at the previous moment, and the two are linearly related, the historical current source of the resistor is always 0, that is, .

[0209] Discretizing the inductance trapezoidal method of formula (1-13), we have:

[0210] (1-14)

[0211] (1-15)

[0212] Arranged

[0213] (1-16)

[0214] in ;

[0215] Then the characteristic equation of the inductor L after discretization using the implicit trapezoidal method is:

[0216] (1-17)

[0217] The calculation formula of the historical current source is

[0218] (1-18)

[0219] like Figure 3 The RL circuit diagram after the trapezoidal method is discretized. Using Kirchhoff's law for node 1, we have

[0220] (1-19)

[0221] Arranged

[0222] (1-20)

[0223] Since the right end of the inductor is grounded, the voltage across the inductor is equal to the voltage at node 1.

[0224] (1-21)

[0225] Substituting (1-21) into (1-18) yields

[0226] (1-22)

[0227] That is, the implicit trapezoidal electromagnetic transient characteristic value of the RL circuit is

[0228] (1-23)

[0229] The relationship between eigenvalues ​​and numerical oscillation phenomena is Figure 3 In the circuit breaker operation, R is equivalent to an infinite value, the resistance is disconnected, and only the inductance L remains in the circuit;

[0230] Backward Euler method eigenvalues ​​and numerical oscillations

[0231] (2-1)

[0232] set up is a constant, from (1-11) we get

[0233] According to the switching characteristics, when hour Substituting into formula (1-5) we get

[0234] (2-2)

[0235] From the above formula, we can see that the voltage across the inductor decays to 0 quickly and no numerical oscillation occurs. Therefore, the essence of the circuit breaker operator not generating numerical oscillation under the backward Euler method discretization is that the characteristics of the update matrix of the historical current source are: , the historical current source decays rapidly to 0.

[0236] This essence can also be seen in the circuit diagram of the discretized current source. Figure 4Discrete circuit diagram after R is disconnected. The resistor is disconnected, and only the inductor L is in the circuit. At this time, the only power source in the circuit is the historical current source. According to Ohm's law,

[0237] (2-3)

[0238] Implicit Trapezoidal Method Eigenvalues ​​and Numerical Oscillations

[0239] set up is a constant, from (1-23) we get

[0240] (2-4)

[0241] According to the switching characteristics, when hour Substituting into formula (1-17) we get

[0242] (2-5)

[0243] From the above formula we can see will be The amplitude swings up and down, generating numerical oscillation. The essence of the circuit breaker operator not generating numerical oscillation under the post-implicit trapezoidal discretization method is: the characteristics of the update matrix of the historical current source , the direction of the historical current source is It is a periodic oscillation change.

[0244] Similarly, this essence can also be seen from the circuit diagram that carries the historical current source after discretization. The resistor is disconnected and there is only inductance L in the circuit. Figure 4 This is the discrete circuit diagram after R is disconnected. At this time, there is only one power source in the circuit, the historical current source. According to Ohm's law:

[0245] (2-6)

[0246] Determining the number of steps used in the backward Euler method when numerical oscillation occurs

[0247] Install the circuit breaker Operation at all times, the initial value of the historical current source is , generating numerical oscillations. The electromagnetic transient simulation calculation is switched to the backward Euler method.

[0248] From equations (2-4) and (2-16), we know that when the step size of the backward Euler method is When the equivalent conductance of the backward Euler method and the step size of the trapezoidal method are Therefore, to simplify programming, when a disturbance occurs in the network, the simulation step size can be reduced to the original , and then calculate the historical current source of each component using the backward Euler method. After that, the calculation is continued by the trapezoidal method. Since the step size is reduced by half, the accuracy of the backward Euler method is improved, and the calculation is quickly switched back to the trapezoidal method, so the error caused by the backward Euler method can be ignored.

[0249] For distributed parameter transmission lines, the Berelon equivalent circuit shows that its equivalent conductance is independent of the step size, so there is no need to modify its equivalent conductance after using the backward Euler method.

[0250] Assume use k After using the half-step backward Euler method, the numerical oscillation is effectively suppressed.

[0251] The historical current source value after the first half-step backward Euler method is known from equation (2-10):

[0252] (3-1)

[0253] The historical current source value after the second half-step backward Euler method is

[0254] (3-2)

[0255] After k-step backward Euler method calculation, the historical current source value is

[0256] (3-3)

[0257] when Decay to of times, it can be considered , from (2-3) we know that the numerical oscillation has been effectively suppressed at this time.

[0258] Right now

[0259] (3-4)

[0260] but

[0261] (3-5)

[0262] Taking the base 10 logarithm of both sides, we get

[0263] (3-6)

[0264] but

[0265] (3-7)

[0266] The minimum number of steps is:

[0267] (3-8)

[0268] Oscillation voltage after k-step half-step backward Euler method calculation

[0269] (3-9)

[0270] (3-10)

[0271] Then proceed The numerical oscillation voltage after the half-step backward Euler method is

[0272] (3-11)

[0273] From formula (3-11), we can see that After the step-backward Euler method, the numerical oscillation voltage can be suppressed to approximately 0, and the accuracy of the simulation calculation is minimally affected.

[0274] In this example, the circuit parameters and calculation parameters are shown in Table 1.

[0275] Table 1 Circuit parameters and calculation parameters

[0276]

[0277] The calculation results using the implicit trapezoidal method, as analyzed above, produce the amplitude , the frequency is The abnormal numerical oscillation seriously affects the calculation results. Figure 5 is the voltage across the inductor calculated using the implicit trapezoidal method; Figure 6 is the inductor historical current source value calculated using the implicit trapezoidal method.

[0278] Calculation results using traditional numerical oscillation suppression methods

[0279] The traditional critical damping adjustment (CDA) method is used to suppress numerical oscillations, that is, after the circuit breaker S is operated, the backward Euler method is used for two and a half steps and then the trapezoidal method is used for calculation. Figure 7 and Figure 8 It can be seen that using this method, the voltage oscillation amplitude is suppressed to a certain extent, but not completely. After exiting the backward Euler method, the voltage value across the inductor and the historical current source value continue to oscillate.

[0280] Calculation results using the method proposed in this patent

[0281] Substituting the data in Table 1 into formula (3-8), we can see that the number of steps of the half-step backward Euler method should be:

[0282]

[0283] from Figure 9 and Figure 10 The results show that the use of this patented method can suppress the voltage across the inductor and the historical current source to approximately 0, effectively suppressing the numerical oscillation phenomenon and minimizing the impact on calculation accuracy.

[0284] In summary: The results show that this patented method can more effectively suppress the numerical oscillation phenomenon in electromagnetic transient simulation by accurately judging the number of steps to use the backward Euler method, and can avoid the impact of simulation calculation accuracy caused by excessive use of the backward Euler method.

[0285] The foregoing description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein and should not be construed as excluding other embodiments. Rather, the present invention can be used in various other combinations, modifications, and environments and can be modified within the scope of the concept described herein through the above teachings or techniques or knowledge in the relevant field. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention are intended to be protected by the appended claims.

Claims

1. A numerical oscillation suppression method for electromagnetic transient simulation based on eigenvalue analysis, characterized in that: The following steps are involved: Step 1: The electromagnetic transient model discretizes the characteristic formulas of the components with memory in the circuit into linear conductance and historical current sources. During normal simulation calculations, the implicit trapezoidal method is used to calculate the state variables and eigenvalues ​​of the electromagnetic transient model. The circuit state variables and the eigenvalues ​​of the electromagnetic transient model at the previous moment determine the value of the historical current source. Step 2: After the circuit breaker is operated, the electromagnetic transient characteristic value of the circuit is calculated under the backward Euler method discretization, and the half-step backward Euler method calculation of k steps that completely suppresses the numerical oscillation phenomenon is obtained based on the electromagnetic transient characteristic value; the simulation step size is reduced to the original simulation step size. , calculate the historical current source of each component according to the backward Euler method, where the linear conductance value remains unchanged, and perform k steps of half-step backward Euler method calculation; Step 3: After performing k-step half-step backward Euler method calculations, switch back to the implicit trapezoidal method to continue the calculations and complete the numerical oscillation suppression of the magnetic transient simulation; After the circuit breaker is operated, the electromagnetic transient characteristic value of the circuit is calculated under the backward Euler method discretization, and the half-step backward Euler method calculation of k steps that completely suppresses the numerical oscillation phenomenon is obtained based on the electromagnetic transient characteristic value; the simulation step size is reduced to the original simulation step size , calculate the historical current source of each component according to the backward Euler method, where the linear conductance value remains unchanged, and perform k steps of half-step backward Euler method calculation, including: If used After the half-step backward Euler method, the numerical oscillation is effectively suppressed. The historical current source value after the first half-step backward Euler method is: The historical current source value after the second half-step backward Euler method is: After performing the k-step backward Euler method calculation, the historical current source value is: when Decay to of times, ,Right now but get but The minimum number of steps for: The oscillating voltage after k steps of half-step backward Euler method calculation is: Then proceed The numerical oscillation voltage after the half-step backward Euler method is: 。 2. The method for suppressing numerical oscillations in electromagnetic transient simulation based on eigenvalue analysis according to claim 1, characterized in that: The circuit contains memory elements, which include inductors and capacitors.

3. The method for suppressing numerical oscillations in electromagnetic transient simulation based on eigenvalue analysis according to claim 1, characterized in that: During normal simulation calculations, the implicit trapezoidal method is used to calculate the eigenvalues ​​of the electromagnetic transient model. After the circuit breaker is operated, the backward Euler method is used to calculate the eigenvalues ​​of the electromagnetic transient model. The circuit state variables and the characteristic values ​​of the electromagnetic transient model at the previous moment determine the value of the historical current source, including: For RL circuit, for the resistor R have For inductors L have For inductance L Trapezoidal discretization, get in Then the characteristic equation of the inductor L after discretization using the implicit trapezoidal method is: The calculation formula of the historical current source is: Then there is get but get Then the implicit trapezoidal electromagnetic transient characteristic value of the RL circuit is: Eigenvalues ​​under the backward Euler discretization For RL circuits For resistors R have For inductors L have Discretization using the Backward Euler method Arranged in ; Then the characteristic equation of the inductor L after discretization using the backward Euler method is: in have Arranged but get That is, the backward Euler method electromagnetic transient eigenvalue of the RL circuit is: 。 4. The electromagnetic transient simulation numerical oscillation suppression system based on eigenvalue analysis is characterized by: The electromagnetic transient simulation numerical oscillation suppression method based on eigenvalue analysis described in any one of claims 1-3 includes an electromagnetic transient simulation module, a data processing module and a data storage module; the electromagnetic transient simulation module and the data storage module are respectively connected to the data processing module.

Images