Power system time domain simulation method based on Jacobian matrix adaptive updating

Abstract

The invention provides a power system time domain simulation method based on Jacobian matrix adaptive updating. Simulation performance is improved under an alternate solution framework. Firstly, system variables are initialized, and a numerical integration method and a nonlinear algebraic equation set iteration format are determined; in the simulation process, comparing the size relation between the number of iterations of each simulation time step and the updating threshold value of the Jacobian matrix, and judging whether the Jacobian matrix needs to be updated or not; solving a state variable and a network algebraic variable based on a nonlinear algebraic equation set iteration format and a network equation; after the calculation of the current time step is finished, adjusting a jacobian matrix updating threshold according to the actual iteration times of the current time step; and judging whether the end time is reached or not, and selecting to execute the next time step simulation or end the calculation. According to the method, by designing a threshold-driven Jacobian matrix updating strategy, the matrix updating frequency is flexibly adjusted, the calculation speed is increased while convergence is guaranteed, and the simulation performance requirement of a power system can be met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system simulation, and in particular to a power system time domain simulation method based on adaptive updating of a Jacobian matrix. BACKGROUND

[0002] With the continuous construction of new power systems, the penetration rate of new energy equipment and power electronic devices in power systems is rising, and at the same time, the models of various new energy equipment are gradually evolving towards high dimension and mixing, which rapidly expands the dimension of the differential algebraic equation set of power system transient simulation. In addition, the influx of a large number of power electronic devices and controllers makes the dynamic characteristics of variables in the model more obvious, and the rigidity problem of the system is more prominent, further affecting the selection of simulation algorithm and simulation time step. The above problems bring problems in convergence and solution speed to the simulation solving algorithm. In order to ensure numerical stability in the simulation process, the implicit trapezoidal method with A-stability, relatively high simulation accuracy and simple and clear principle is usually used for numerical integration in the face of simulation calculation demand of rigid power system.

[0003] Under the numerical integration solving framework, the power system differential algebraic equation set can be differentiated into a nonlinear algebraic equation set for solving. The existing solving methods of nonlinear algebraic equation set include simple iteration method, Newton method, etc. For large-scale power system simulation, Newton method needs to solve the inverse matrix of large-scale Jacobian matrix, and the calculation amount is very large, which is limited in practical simulation. For this, part of the research simplifies the Newton method to obtain the VDHN method with lower calculation amount, but the convergence of VDHN is more and more difficult to meet the demand under the condition that the system dynamic process is more and more complex. The simple iteration method has low calculation amount and is favored by many simulation software and research literature, but its convergence performance is greatly affected by the simulation time step, and when the convergence is low, the time step needs to be adjusted to ensure convergence, which limits the improvement of simulation performance.

[0004] The above methods have certain deficiencies in solving large-scale system simulation calculation. Therefore, it is urgent to optimize the simulation solving method to improve the efficiency of simulation iteration as much as possible while ensuring convergence. SUMMARY

[0005] The purpose of the present application is to provide a power system time domain simulation method based on adaptive updating of a Jacobian matrix to solve the problems in the prior art.

[0006] The technical solution of the present application is: a power system time domain simulation method based on adaptive updating of a Jacobian matrix, comprising the following steps: S1, initialization: initializing system state variables , network algebraic variables Determine the numerical integration algorithm and the algorithm for solving nonlinear algebraic equations; Given the Jacobian matrix, update the first threshold. Second update threshold The initial value; given The step size of the reduction Adjust step size Lower the trigger threshold and raising the trigger judgment threshold Given The step size of the reduction Adjust step size Adjusting the trigger threshold based on the number of historical iterations The current total number of iterations determines the trigger threshold for lowering. and raising the trigger judgment threshold Given simulation time steps and variable iteration convergence threshold Set the initial state of the Jacobian matrix update judgment flag to update; The numerical integration algorithm is determined, and the calculation formula of the numerical integration algorithm is as follows: ; in, , The first The state variable vector and network algebra variable vector at each time step; , The first The state variable vector and network algebra variable vector at each time step; State variables Regarding time The derivative function; The formula for the nonlinear algebraic equation system solving algorithm used in step S6 for updating the state variables is as follows:

[0007] in, , The vector consists of the state variables of the k-th and k+1-th iterations at the (n+1)-th time step, respectively; This is a vector composed of the network algebraic variables of the (n+1)th time step and the kth iteration; This represents the residual corresponding to the state variable vector at the (n+1)th time step and the kth iteration. This represents the Jacobian matrix corresponding to the state variables in the m-th iteration at time step l; The calculation formula is: The determination of l and m is as follows: the initial values ​​of l and m are both set to 0; each time step S6 is executed, the Jacobian matrix update judgment flag is read. If the state of the Jacobian matrix update judgment flag is "updated", that is, the current iteration calculation needs to update the Jacobian matrix, and then... , The values ​​are updated to respectively , If the Jacobian matrix update check flag is set to "not update", then no adjustment is made. l and m That is, using historical Jacobian matrix information; S2, Update simulation time ; S3, Number of Update Iterations If the step executed before S3 is S2, then If the step executed before S3 is S8, then Where k is the current iteration number, This represents the updated iteration count. Then, set the value of the current iteration number k to... The value;

[0008] S4. Calculation of State Variable Residuals: Calculate the residuals corresponding to the state variable vectors. ,in For the first n +1 hour step k The vector consisting of the state variables from +1 iterations; For the first n +1 hour step k The vector composed of network algebra variables from each iteration; S5. Jacobian Matrix Update Decision: Based on the current iteration count. Total number of iterations in the previous time step Jacobian matrix first update threshold And Jacobian matrix second update threshold Set an update flag for the Jacobian matrix; if the condition is met... or If the condition is met, the state of the Jacobian matrix update judgment flag is set to update; if the condition is not met, the state of the Jacobian matrix update judgment flag is set to no update. S6. State Variable Update: Read the Jacobian matrix update judgment flag; if the flag indicates an update, then update the Jacobian matrix. ,make , ,in Indicates the first l Time Step mThe Jacobian matrix corresponding to the state variables in the next iteration; if the flag's state is not updated, then and Keeping it unchanged, use the historical Jacobian matrix; Execute ; S7. Network Algebra Variable Update: Update network algebra variables. The update method is as follows:

[0009] in, Inject current calculation functions into nodes. The network admittance matrix of the system. For the updated version n +1 hour step k+1 The vector composed of network algebra variables from each iteration; S8. Alternating Iteration Convergence Judgment: Judgment and Does it meet the requirements? If satisfied, update the total number of iterations in the previous step. The update method is that the values ​​of the current simulation time and the simulation time step satisfy the following conditions: Then take If not satisfied Then The value is set to the currently recorded value. The value of . Update complete. Then, update the total number of iterations at the current time step. And continue executing S9;

[0010] Otherwise, return to S3; This represents the infinite norm of the vector *. S9. Algorithm parameter adaptive adjustment: Based on S8 Total number of iterations in the previous step ,Adjustment and : Update as follows: , Update as follows: , Will and The value assigned and ; S10, Time Domain Simulation End Judgment: Determine whether the current simulation time has reached the simulation end time given in S1; if it has, the simulation ends; otherwise, return to S2.

[0011] Specifically, in step S4, the state variable vector corresponds to the residual. The calculation method is as follows: .

[0012] Specifically, in step S2, the method for updating the simulation time is as follows: Estimated simulation time for normal updates The calculation formula is: ,in The current simulation time; determine to Does the time period include a discrete event occurring or ending? If not, then... Set to the current simulation time; if it exists, adjust the simulation time step using the following formula: ,in The earliest discrete event time within this time period. This is the adjusted simulation time step. (The result is...) Then, recalculate. The calculation formula is: ,get Then, set the current simulation time. The value is updated to .

[0013] Compared with the prior art, the present invention has the following beneficial effects: 1) This invention avoids recalculating and solving large-scale Jacobian matrices in each iteration by adaptively determining whether to update the Jacobian matrix, which significantly reduces the amount of computation and thus speeds up the overall computation speed of time-domain simulation. 2) This aspect uses an adaptive update strategy for the Jacobian matrix to dynamically adjust the update threshold according to the actual convergence situation in the simulation process. This allows the Jacobian matrix to be updated more frequently when convergence is difficult to enhance convergence, and the updates to be reduced when convergence is stable to improve efficiency, thus achieving a balance between efficiency and stability. 3) For complex rigid systems containing a large number of new energy devices and power electronic devices, the adaptive strategy of this invention can better cope with the differences in dynamic characteristics and rigidity problems, thus improving the robustness of the algorithm; 4) Compared to the simple iterative method, the adaptive update strategy of this invention reduces the sensitivity to the selection of simulation time steps to a certain extent, ensuring convergence without frequent adjustments to the time steps, simplifying simulation settings and reducing dependence on simulation time steps; compared to the commonly used extremely dishonest Newton method (VDHN method), this invention designs the number of iterations within the time step and the first update threshold. VDHNThe size judgment logic between 1 increases the number of Jacobian matrix update operations allowed within a single time step, which helps improve the convergence of the overall iterative calculation, avoids frequent time step callbacks due to convergence failure, and further supports the improvement of simulation efficiency.

[0014] 5) The method proposed in this invention has good flexibility. By modifying the adjustment step size in the Jacobian matrix update threshold adjustment strategy, it can take into account the computational needs of different simulation scenarios. Attached Figure Description

[0015] To more clearly illustrate the technical solution of the present invention, the accompanying drawings are briefly described below: Figure 1 This is a schematic diagram of the time-domain simulation method for power systems based on adaptive update of the Jacobian matrix provided in an embodiment of the present invention; Figure 2 This is a topology diagram of the test system in the power system time-domain simulation method based on adaptive update of the Jacobian matrix provided in the embodiments of the present invention; Figure 3 The waveform diagram shows the test results of the power system time-domain simulation method based on adaptive update of Jacobian matrix provided in the embodiments of the present invention. Detailed Implementation

[0016] 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, 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 scope of protection of the present invention.

[0017] Example: A time-domain simulation method for power systems based on adaptive updating of the Jacobian matrix, such as Figure 1 As shown, it includes the following steps: S1. Initialize system variables, simulation algorithm, algorithm parameter configuration, and simulation parameter configuration.

[0018] The initialization of system variables includes the initial values ​​of given state variables and network algebra variables. , .

[0019] The initialization simulation algorithm includes a deterministic numerical integration algorithm and an algorithm for solving nonlinear algebraic equations.

[0020] The initialization algorithm parameter configuration includes a first update threshold for the given Jacobian matrix. VDHN 1 and second update threshold VDHNInitial value of 2; given VDHN The downward adjustment step size Δ1 and the upward adjustment step size Δ2 are 1. VDHN 2. The downward adjustment step size Δ3 and the upward adjustment step size Δ4; given VDHN 1. Lowering the trigger threshold k 1 and raising the trigger threshold k 2; given VDHN 2. Adjusting the trigger threshold based on the number of historical iterations k 3. Adjust the trigger threshold based on the current total number of iterations. k 5 and raising the trigger judgment threshold k 4. The initial state of the Jacobian matrix update judgment flag is given. The Jacobian matrix update judgment flag is the basis for whether the Jacobian matrix is ​​updated when updating the state variables in step S8, including two states: "update" and "do not update"; the initial state of the Jacobian matrix update judgment flag is set to "update".

[0021] The initialization simulation parameter configuration includes a given simulation end time, simulation time step, variable iteration convergence threshold, and discrete event information. The discrete event information includes discrete event type, discrete event occurrence location, discrete event occurrence time, discrete event end time, and discrete event parameter configuration.

[0022] The numerical integration algorithm is determined, and the calculation formula of the numerical integration algorithm is as follows:

[0023] in, x n+1 , x n The vector consists of the state variables at time step (n+1) and time step (n), respectively. u n+1 , u n The vector consists of the network algebraic variables at time step (n+1) and time step (n), respectively; the subscripts indicate the simulation time. f State variables x Regarding time t The derivative function; h This is for simulation timing.

[0024] The algorithm for solving nonlinear algebraic equation systems is determined, and the calculation formula for the algorithm is as follows:

[0025] in, , The first n +1 hour step k The next iteration and kThe vector composed of state variables at the +1-th iteration; For the n +1-th time step and the k vector composed of network algebraic variables at the -th iteration; Denote the vector composed of the residuals corresponding to the state variables at the (n + 1)-th time step and the -th iteration; k Denote the l time step and the m Jacobian matrix corresponding to the state variables at the -th iteration. And Their calculation formulas are as follows:

[0026]

[0027]

[0028] In the calculation formulas of the non - linear algebraic equation solving algorithm, l And m are determined as follows: l And m Their initial values are both set to 0.

[0029] Each time step S6 is executed, read the Jacobian matrix update judgment flag. If the status of the Jacobian matrix update judgment flag is "update", that is, the current iterative calculation needs to update the Jacobian matrix, update l And m in the following way:

[0030] If the status of the Jacobian matrix update judgment flag is "not update", then do not adjust l And m , that is, adopt the historical Jacobian matrix information. [[ID=5�]]

[0031] In one embodiment, taking a 39 - node simulation system as an example, the system includes 10 synchronous generators; 2 photovoltaic devices; 3 wind turbine devices. The rated frequency of the system is 50Hz. The test system topology diagram is as Figure 2 shown, [[ID=6۰]] Figure 2 where "light" represents photovoltaic; "wind" represents wind turbine, "G" represents synchronous generator, " " represents node load. The node parameters of the test system are shown in Table 1, and the branch parameters are shown in Table 2. According to the system parameters in Table 1 and Table 2, the network admittance matrix of the system can be directly constructed.

[0032] Table 1 Node parameter table of the test system

[0033] Table 2 Test System Branch Parameter Table

[0034] In one embodiment, VDHN 1 and VDHN 2. The initial value is given as 0. Δ1 and Δ2 are set to 1 and 1 respectively, and Δ3 and Δ4 are set to 3 and 2 respectively; k 1. k 2. k 3. k 4. k The values ​​5 are set to 15, 8, 5, 3, and 15 respectively; the simulation end time is set to 10 s; the simulation time step is set to 0.001 s; and the variable iteration convergence threshold is set to... Discrete event information is shown in Table 3:

[0035] Table 3 Discrete Event Information Table

[0036] S2. Update simulation time. Determine the current simulation time before solving the system equations at each time step.

[0037] The specific method for updating simulation time is as follows: First, estimate the simulation time for normal updates. T new :

[0038] in, T new The estimated simulation time after the update; T now The original simulation time, h The simulation time step is given in step S1.

[0039] Get the estimated update time T new Then, based on the simulation event information given in step S1, the simulation time is determined. T now ~ T new Does the time period contain any discrete events that occur or end? If not, T new Set to the current simulation time; otherwise, adjust the simulation time step to match the current simulation time. T new The specific method for corresponding discrete event occurrence times is as follows:

[0040]

[0041] in, Tevent for T now ~ T new The time corresponding to the occurrence or end of a discrete event within a time period, if T now ~ T new Within a time period, if there are multiple discrete events occurring or ending, the minimum time corresponding to all these scenarios is taken as the minimum time. T event ; h event This is the adjusted simulation time step. (The result is...) h event Then, the updated simulation time will be... T new Set as T now + h event .

[0042] S3. Update the iteration count. Before the iteration calculation begins, determine the iteration count at the current time step. The method for updating the iteration count is as follows: if the step executed before the current step S3 is S2, then update the current iteration count. k Set to 0; if the step executed before the current step S3 was S8, then update the current iteration number according to the following formula:

[0043]

[0044] in k This represents the current iteration number. k new This represents the updated iteration count. k new Then, the current iteration number will be... k The value is set to k new The value of .

[0045] S4. Calculation of State Variable Residuals. Calculate the state variable residuals according to the solution format of the numerical integration algorithm. . The calculation method is as follows:

[0046]

[0047] S5. Jacobian matrix update judgment. Based on the current iteration number. k Number of iterations in the previous time step iter n Jacobian matrix first update threshold VDHN 1 and the second update threshold of the Jacobian matrixVDHN 2. Determine whether the current iteration process requires an update operation of the Jacobian matrix. The specific method is as follows:

[0048] If satisfied If the condition is met, it is determined that a Jacobian matrix update is required, and the status of the Jacobian matrix update judgment flag is set to "update"; if the above condition is not met, it is determined that a Jacobian matrix update is not required, and the status of the Jacobian matrix update judgment flag is set to "do not update".

[0049] S6. State Variable Update. Update the state variables according to the nonlinear algebraic equations solving algorithm determined in step S1. The specific update method is as follows:

[0050]

[0051] in, The updated result is the first n +1 hour step k The vector consisting of the state variables corresponding to +1 iterations; For the first n +1 hour step k The vector composed of the state variables of each iteration; For the first n +1 hour step k The vector composed of network algebra variables from each iteration; For the (n+1)th time step obtained in step S4, the... k The vector formed by the residuals corresponding to the state variables in each iteration; Indicates the first l Time Step m The Jacobian matrix corresponding to the state variables in the next iteration. l and m The method for determining it is as follows: l and m The initial values ​​of all are set to 0. When executing step S6, the Jacobian matrix update judgment flag is read. If the status of the Jacobian matrix update judgment flag is "updating," meaning the current iteration calculation requires updating the Jacobian matrix, it is updated as follows: l and m :

[0052]

[0053] If the Jacobian matrix update judgment flag is in the state of "no update", then no adjustment is made. l and m That is, using historical information from the Jacobian matrix.

[0054] S7. Solving for network algebraic variables. The method for updating network algebraic variables is as follows:

[0055]

[0056] in, i Inject a current calculation function into a given node. Y The network admittance matrix of the system. This is the vector composed of the network algebraic variables obtained at the (n+1)th time step and the (k+1)th iteration after the update.

[0057] S8. Convergence Judgment of Alternating Iteration of System Equations. Based on the differences between the current state variables and network algebra variables and those of the previous iteration, determine whether the alternating iterative solution at the current time step has converged. The specific method is as follows:

[0058] If the state variable vectors of the (k+1)th and (k)th iterations at the current time step (n+1) are... , and the network algebraic variable vectors at the (k+1)th and kth iterations at the current time step. , satisfy:

[0059] If the iterative calculation has converged, proceed to the next step; otherwise, return to step S3. The variable iteration convergence threshold given in step S1; This indicates the calculation of the infinite norm of the vector *.

[0060] S9. Adaptive adjustment of simulation algorithm parameters. Adjust the first update threshold of the Jacobian matrix based on the number of iterations required for convergence in this time step. VDHN 1 and second update threshold VDHN 2. The adjustment method for the Jacobian matrix update threshold is as follows:

[0061]

[0062]

[0063] Where Δ1 and Δ2 are given in step S1 VDHN The downward and upward adjustment steps are 1, and Δ3 and Δ4 are given in step S1. VDHN 2. The downward and upward adjustment steps; k 1. k 2 is given in step S1 VDHN 1. The threshold for triggering a decrease and the threshold for triggering an increase; k 3. k 5 is the value given in step S1. VDHN 2. The historical iteration count is used to determine the trigger threshold for lowering the threshold, and the current total iteration count is used to determine the trigger threshold for lowering the threshold.k 4 is the value given in step S1. iter The threshold for triggering the judgment of 2 is increased. This represents the total number of iterations at the current time step. VDHN n This represents the total number of iterations in the previous time step. VDHN 1new , VDHN 2new These are the first and second update thresholds for the adjusted Jacobian matrix. After updating according to the above method, [the following will be used]. VDHN 1new , VDHN 2new The value assigned VDHN 1 and Figure 3 2.

[0064] S10. Time-domain simulation end judgment. After the system equations at the current time step are solved, confirm whether the simulation time has reached the simulation end requirement. Specifically, if the current simulation time reaches the simulation end time given in step S1, the simulation ends; otherwise, return to step S2.

[0065] In one embodiment, the simulation time obtained by the power system time-domain simulation method based on adaptive update of the Jacobian matrix proposed in this invention (denoted as the algorithm of this invention) is shown in Table 3. To better verify the effectiveness of the algorithm of this invention, based on the implicit trapezoidal rule method for numerical integration, the simulation results of Newton's method (denoted as NR method), simple iterative method, and the extremely dishonest Newton method (denoted as VDHN method) commonly used in existing research are further compared to analyze the performance and accuracy of the method of this invention. The simulation time of various algorithms is shown in Table 4. Table 4 Comparison of Algorithm Time

[0066] In one embodiment, the simulation waveform results using the algorithm of this invention, the NR method, the simple iterative method, and the VDHN method are as follows: Figure 3 As shown.

[0067] Comparison Table 4 and ​ It can be seen that the power system time-domain simulation method based on adaptive update of Jacobian matrix designed in this invention has a shorter simulation time compared with the other three types of algorithms, and the efficiency of alternating iterative solution is improved. At the same time, the accuracy of the simulation results corresponding to the algorithm is basically consistent with the other three types of algorithms, which is conducive to meeting the needs of power system simulation analysis.

[0068] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined in this invention may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A power system time domain simulation method based on adaptive update of Jacobian matrix, characterized in that, The method comprises the following steps: S1, initialization: initialize system state variables , network algebraic variables ; Determine the numerical integration algorithm and the algorithm for solving nonlinear algebraic equations; given the first update threshold of the Jacobian matrix. Second update threshold The initial value; given The step size of the reduction Adjust step size Lower the trigger threshold and raising the trigger judgment threshold ; Given Downward step size Upward step size Historical iteration number judgment downward trigger threshold Current total iteration number judgment downward trigger threshold And upward trigger judgment threshold ; Given simulation time step And variable iteration convergence threshold ; Set the initial state of the Jacobian matrix update judgment flag to update; The formula of the numerical integration algorithm is: ； in, , The first The state variable vector and network algebra variable vector at each time step; , The first The state variable vector and network algebra variable vector at each time step; State variables Regarding time The derivative function; The formula of the nonlinear algebraic equation solving algorithm used in the state variable updating in step S6 is: ； wherein, , are vectors of state variables at the (n+1)th time step for the kth and k+1th iterations, respectively; is a vector of network algebraic variables at the (n+1)th time step for the kth iteration; denotes the corresponding residual for the kth iteration state variable vector at the (n+1)th time step; denotes the Jacobian matrix corresponding to the mth iteration state variable at the lth time step; The calculation formula is: The determination of l and m is as follows: the initial values ​​of l and m are both set to 0; each time step S6 is executed, the Jacobian matrix update judgment flag is read. If the state of the Jacobian matrix update judgment flag is "updated", that is, the current iteration calculation needs to update the Jacobian matrix, and then... , The values ​​are updated to respectively , If the Jacobian matrix update check flag is set to "not update", then no adjustment is made. l and m That is, using historical Jacobian matrix information; S2, update simulation time ; S3, update the iteration number If the step executed before S3 is S2, then If the step executed before S3 is S8, then where k is the current iteration number, is the updated iteration number. The value of k is set to after the iteration number is updated. the value of k. S4, state variable residual calculation: calculate the residual corresponding to the state variable vector wherein is the vector of state variables at the n +1 time step of the k +1 iteration; is the vector of network algebraic variables at the n +1 time step of the k +1 iteration; S5, Jacobian matrix update judgment: based on the current time step iteration number , the total iteration number of the last time step , the first update threshold of the Jacobian matrix , and the second update threshold of the Jacobian matrix , set the update judgment flag of the Jacobian matrix, if the condition or is met, set the state of the Jacobian matrix update judgment flag to update; if the above condition is not met, set the state of the Jacobian matrix update judgment flag to not update; S6. State Variable Update: Read the Jacobian matrix update judgment flag; if the flag indicates an update, then update the Jacobian matrix. ,make , ,in Indicates the first l Time Step m The Jacobian matrix corresponding to the state variables in the next iteration; if the flag's state is not updated, then and Keeping it unchanged, use the historical Jacobian matrix; Execute ; S7, network algebra variable updating: updating the network algebra variable The updating method is: ； wherein, a function for injecting current into the nodes, a conductance matrix for the system network, a vector of network algebraic variables at the updated n +1 time step, k+1 the updated network algebraic variables at the +1 time step. S8, alternately iterative convergence judgment: judging whether and is satisfied; If yes, update the total iteration number of the last time step , the update mode is that the current simulation time and the simulation time step value meet , the value of is taken , if not, the value of is set to the value of the currently recorded , after updating , update the total iteration number of the current time step , and continue to execute S9; Otherwise, go to S3; denotes the infinity norm of the vector *; S9, algorithm parameter self-adaptive adjustment: according to S8 obtained and the total iteration number of the last time step , adjust and . The adjustment method is as follows: The following formula is used for the calculation : ， The following formula is used for the calculation : ， The updated values of and are and respectively. After the calculation, the values of and are assigned to and respectively. S10, time domain simulation end judgment: judging whether the current simulation time reaches the given simulation end time in S1; if yes, the simulation ends; otherwise, returning to S2.

2. The method of claim 1, wherein, In step S4, the state variable vector corresponds to the residual The calculation method is as follows: .

3. The method of claim 1, wherein, In step S2, the method for updating the simulation time is: estimated normal update simulation time , the calculation formula is , wherein is the current time; determining to whether there is a discrete event occurrence or a discrete event end within the time period: if not, then is set to the current simulation time; If present, adjust the simulation time step, the adjustment formula is where is the earliest occurrence of a discrete event time within the time period, is the adjusted simulation time step, resulting in After that, recalculate , the calculation formula is and update the value of the current simulation time step to , Get After that, the value of the current simulation time is updated to .

Images