Robustness analysis and optimization method of independent power system for MBSE

Abstract

The present invention discloses a robustness analysis and optimization method for an independent power system applicable to MBSE, comprising the following steps: 1. Establishing a simulation-based robustness analysis process for an independent power system, including: 101. Rated design; 102. Sensitivity analysis; 103. Statistical analysis; 104. Stress analysis; and 105. Failure mode analysis; 2. Conducting a robustness assessment of the independent power system based on robustness, including: 201. Determining an independent power system robustness assessment indicator; 202. Determining an independent power system robustness assessment method; and 203. Evaluating the robustness of the independent power system; 3. Conducting robust design optimization of the independent power system based on Pareto analysis, including: 301. Analyzing independent power system parameters based on Pareto analysis; and 302. Determining parameters for robust optimization design of the independent power system. The present invention can optimize the system performance of a given design and enhance the robustness of the independent power system.

Description

Technical Field

[0001] The present invention belongs to the technical field of independent power system design, and in particular relates to an independent power system robustness analysis and optimization method suitable for MBSE. Background Art

[0002] Independent power systems are not only a form of electricity consumption for industrial products such as aircraft and automobiles, but also an important model for future ground power grids to face the growing demand for electricity and the development and utilization of new energy.

[0003] The traditional R&D process for independent power systems involves first designing, manufacturing, and testing prototypes of each component, then building a comprehensive system test platform. Using extensive physical test data, the system's performance is verified to meet requirements. If all requirements are not met, component designs are then redesigned, parameters modified, and retested, with continuous iterations until all requirements are fully met. This "whack-a-mole" R&D process is characterized by long development cycles and high costs. Furthermore, once a design is solidified, designers are reluctant to make modifications or adjustments.

[0004] After an independent power system is developed and put into operation, it is subject to not only minor disturbances such as aging, wear, and changes in environmental and operating conditions, but also major disturbances such as high-power load shedding and system failures. Because laboratory resources cannot fully support verification under these disturbance conditions, especially various fault conditions, a system that previously operated normally in the lab may not meet design requirements under certain real-world conditions.

[0005] To address these issues, robust design methods based on component prototype testing (such as the Taguchi method) can be used to optimize the design of each component in the system during the component design phase. The goal is to ensure that the performance of the independent power system is immune to changes in design technology, component parameters, manufacturing processes, aging, and operating conditions. The designed system can meet the expected performance standards and ensure safe system operation to a greater extent. At the same time, the simplest and most cost-effective design is achieved to avoid over-design of the system.

[0006] However, the traditional robust design method based on component prototype experiments for robust design and optimization of independent power systems still has the following shortcomings:

[0007] (1) For any component, when there are many design factors and factor levels that need to be verified, the number of prototypes required is large. If iterative design is required on top of that, the R&D cost is high;

[0008] (2) The prototype experiments of each component in the system are carried out separately, that is, the robust design and optimization of each component are carried out independently. After the components are connected to form the system, the robustness of the entire independent power system cannot be guaranteed;

[0009] (3) The performance of the independent power system under large disturbances cannot be guaranteed.

[0010] Model-based systems engineering (MBSE) is a key approach to the future digitalization of industrial R&D. MBSE supports the application of formalized modeling for system requirements, design, analysis, verification, and validation, starting from the conceptual design phase and continuing throughout development and later lifecycle phases. Based on model simulation results, designers can understand the operation of independent power systems under different operating conditions and even make many critical decisions (such as component parameters, protection circuit thresholds, and control parameter settings). Currently, there is a lack of independent power system robustness analysis and optimization methods suitable for MBSE. Summary of the Invention

[0011] The technical problem to be solved by the present invention is to address the deficiencies in the above-mentioned prior art and provide an independent power system robustness analysis and optimization method suitable for MBSE, which can optimize the system performance of a given design and improve the robustness of the independent power system.

[0012] To solve the above technical problems, the present invention adopts a technical solution: a robustness analysis and optimization method for an independent power system applicable to MBSE, characterized in that the method comprises the following steps:

[0013] Step 1: Establish a simulation-based independent power system robustness analysis process, including:

[0014] Step 101, rated design: Build a system model of the independent power system, perform time-domain simulation on the model of the independent power system under rated conditions, obtain simulation data of voltage and current, and use the simulation data of voltage and current to analyze two levels of indicators: steady-state characteristic indicators and dynamic characteristic indicators to ensure that the independent power system performs according to the expected performance specifications under rated conditions;

[0015] Step 102, sensitivity analysis: applying small positive and negative disturbances to each parameter of the system model according to the set parameter disturbance ratio d%, performing system simulation and recording the system response dynamic characteristic indicators, calculating the sensitivity per unit value of different parameters relative to different system response dynamic characteristic indicators, until the sensitivity analysis of all parameters in the system model is completed;

[0016] Step 103, statistical analysis: simulate noise factors and perform Monte Carlo simulation based on Latin hypercube sampling on the independent power system model to obtain information on two levels of indicators, namely steady-state characteristic indicators and dynamic characteristic indicators, for the established system model parameters under uncertain noise factors;

[0017] Step 104, stress analysis: searching for voltage and current data of each component from simulation results that meet the steady-state characteristic index requirements in the statistical analysis, obtaining the maximum stress rate, and performing stress analysis based on the maximum stress rate;

[0018] Step 105, Fault Mode Analysis: Model the fault to be analyzed in the system model of the independent power system. By setting components to fail in various ways and at specific times, study how single component failures or combined failure modes affect system performance, and obtain analysis results of two-level indicators of the system under the failure modes.

[0019] Step 2: Conduct an independent power system robustness assessment based on robustness, including:

[0020] Step 201: Determine an independent power system robustness assessment index;

[0021] Step 202: Determine an independent power system robustness assessment method;

[0022] Step 203: Evaluate the robustness of the independent power system according to the independent power system robustness evaluation method determined in step 202;

[0023] Step 3: Perform robust design optimization of the independent power system based on Pareto analysis, including:

[0024] Step 301: Perform independent power system parameter analysis based on Pareto analysis;

[0025] Step 302: Perform parameter determination for robust optimization design of the independent power system.

[0026] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized by: defining the steady-state characteristic index in steps 101 and 103 as a primary index, which is used to reflect the steady-state function of the independent power system and indicates that the current system is in a stable operating state; defining the dynamic characteristic index in steps 101 and 103 as a secondary index, which is used to reflect the dynamic performance of the independent power system and measure the dynamic performance of the system when subjected to different disturbances, so as to select the best system design under the premise of ensuring that the steady-state performance meets the requirements;

[0027] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific process of the rated design in step 101 is:

[0028] Step 1011: Start;

[0029] Step 1012: Build a system model of the target independent power system;

[0030] Step 1013: setting the model parameters to a rated state;

[0031] Step 1014: setting the model simulation working condition to the rated state;

[0032] Step 1015: System time domain simulation to obtain model output voltage and current simulation data;

[0033] Step 1016: extract steady-state characteristic indicators;

[0034] Step 1017: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1018; otherwise, if the steady-state characteristic index does not meet the design requirements, return to step 1013.

[0035] Step 1018: extract dynamic characteristic indicators;

[0036] Step 1019: Determine whether the dynamic characteristic index meets the design requirements. If the dynamic characteristic index meets the design requirements, execute step 10110; otherwise, if the dynamic characteristic index does not meet the design requirements, return to step 1013.

[0037] Step 10110, end.

[0038] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific process of the sensitivity analysis in step 102 is as follows:

[0039] Step 1021: Start;

[0040] Step 1022: Set the system model parameter disturbance ratio d%.

[0041] Step 1023: Select the next system model parameter;

[0042] Step 1024: applying small positive disturbances and negative disturbances to the parameters respectively;

[0043] Step 1025: Perform system simulation to obtain the model output voltage and current response;

[0044] Step 1026: extract dynamic performance indicators;

[0045] Step 1027: Calculate the per-unit positive disturbance sensitivity and the per-unit negative disturbance sensitivity of the parameter relative to different dynamic performance indicators respectively;

[0046] Step 1028: Determine whether the per-unit values ​​of the positive disturbance sensitivity and the per-unit values ​​of the negative disturbance sensitivity of all parameters have been calculated. If the calculation is completed, execute step 1029; otherwise, if the calculation is not completed, return to step 1023.

[0047] Step 1029, end.

[0048] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that: the parameter disturbance ratio d% in step 1022 is less than or equal to 5%;

[0049] The calculation formulas for the per-unit value of the positive disturbance sensitivity and the per-unit value of the negative disturbance sensitivity in step 1027 are: Where Sensitivity′ is the per-unit value of the positive disturbance sensitivity or the per-unit value of the negative disturbance sensitivity, m now is the dynamic performance index value extracted from the sensitivity analysis simulation, m rated is the dynamic performance index value extracted from the rated design.

[0050] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific process of the statistical analysis in step 103 is as follows:

[0051] Step 1031: Start;

[0052] Step 1032: Create a list of key system model parameters to be observed;

[0053] Step 1033: specify distribution and tolerance range for each key parameter;

[0054] Step 1034: Given the estimated accuracy and confidence level, calculate the minimum number of simulations N required for Monte Carlo simulation. min ;

[0055] Step 1035: Perform Latin hypercube sampling on the input parameters within the tolerance range determined in step 1033 to generate N that meet the distribution min Set parameter combination samples; take the initial value of parameter group number i as 1, and the initial value of system stability times n in statistical analysis as 0;

[0056] Step 1036: inject the i-th group of parameters into the system model;

[0057] Step 1037: System time domain simulation to obtain model output voltage and current simulation data;

[0058] Step 1038: extract steady-state characteristic indicators;

[0059] Step 1039: Determine whether the steady-state characteristic index meets the design requirements. If so, execute step 10310. Otherwise, if not, increment the value of i by 1, and return to step 1036.

[0060] Step 10310: Add 1 to the value of the system stability times n in the statistical analysis;

[0061] Step 10311: extract dynamic characteristic indicators;

[0062] Step 10312: Determine whether the value of i is equal to N. min , when the value of i is equal to N min , execute step 10313; otherwise, when the value of i is not equal to N min When , add 1 to the value of i and return to step 1036;

[0063] Step 10313, end.

[0064] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the method of creating the list of key parameters of the system model to be observed in step 1032 adopts one of the following three methods:

[0065] Method A1: Use the results of the sensitivity analysis to select the factors that most affect the change in the performance metric and create a list of key system model parameters to be observed.

[0066] Method A2: Use industry knowledge and rules of thumb to analyze the type of system being analyzed and create a list of key system model parameters to observe;

[0067] Method A3: Rely on the designer's experience and create a list of key parameters of the system model to be observed based on previous designs of similar systems.

[0068] When specifying a distribution for each key parameter in step 1033, the distribution includes normal distribution, average distribution, exponential distribution, and Γ-distribution;

[0069] The method for specifying the tolerance range for each key parameter in step 1033 includes the following methods:

[0070] Method B1: For actual electronic component parameters, query the tolerance range of the parameters from the device data sheet, and estimate the tolerance range based on the system operating temperature and component temperature drift;

[0071] Method B2: The tolerance range of mechanical structure parameters is determined according to the manufacturing tolerance range;

[0072] Method B3: For parameters that need to be measured by specific experiments, the tolerance range is determined based on the measurement error of the instrument;

[0073] Method B4: For parameters obtained by fitting measured data, the fitting error is determined as the parameter tolerance range, or the tolerance range is estimated in combination with the instrument measurement error;

[0074] Given the estimated accuracy and confidence level in step 1034, the minimum number of simulations N required for Monte Carlo simulation is calculated. min When , the calculation formula used is: Where δ is the confidence level and δ∈(0,1), ε is the estimation accuracy and ε∈(0,1);

[0075] When performing Latin hypercube sampling on the input parameters within the tolerance range determined in step 1033 in step 1035, assuming that there are n′ input uncertain parameters in the model, N min Monte Carlo simulations require N extractions in the n′-dimensional vector space. min The specific process of Latin hypercube sampling is:

[0076] Step 10351: Divide each dimensional vector space into N non-overlapping min intervals, so that each interval has the same probability;

[0077] Step 10352: randomly select a point in each interval in each dimensional vector space;

[0078] Step 10353: Randomly extract the points selected in step 10352 from the n′-dimensional vector space to form N min Set vector, the N min The group vector is determined as N for Monte Carlo simulation min Group parameter combination samples.

[0079] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific process of the stress analysis in step 104 is as follows:

[0080] Step 1041: Start;

[0081] Step 1042: Set the value of i to 1;

[0082] Step 1043: Statistically analyze the i-th Monte Carlo simulation;

[0083] Step 1044: Obtain system model Monte Carlo simulation results and device voltage and current data;

[0084] Step 1045: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1046; otherwise, if the steady-state characteristic index does not meet the design requirements, add 1 to the value of i and return to step 1043.

[0085] Step 1046: Calculate the stress rate based on the device voltage and current data;

[0086] Step 1047: Determine whether the value of i is equal to N min , when the value of i is equal to N min , execute step 1048; otherwise, when the value of i is not equal to N min When , add 1 to the value of i and return to step 1043;

[0087] Step 1048: From N min Select the maximum stress rate from the calculated stress rates;

[0088] Step 1049: Perform stress analysis based on the maximum stress rate;

[0089] Step 10410, end.

[0090] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that when calculating the stress rate based on the device voltage and current data in step 1046, the calculation formula used to calculate the stress rate StressRatio(i) during the i-th Monte Carlo simulation is: Among them, max(|E i |) is the maximum value of the device voltage and current in the i-th Monte Carlo simulation, rated(|E i |) is the operating threshold of the device voltage and current set by the designer in the i-th Monte Carlo simulation, U i is the voltage value of the device in the i-th Monte Carlo simulation, I i is the current value of the device in the i-th Monte Carlo simulation.

[0091] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific method of performing stress analysis based on the maximum stress rate in step 1049 is:

[0092] When the maximum stress rate is greater than 1, it means that the voltage and current stresses of the device have exceeded the safe operating limits specified by the designer during stress analysis. The operating threshold of the device needs to be reconsidered. This helps designers discover situations where circuit performance does not meet design goals and guides designers in device selection.

[0093] When the maximum stress rate is ≤1, it means that the device has not reached its stress limit value in the stress analysis, and the device selection is reasonable;

[0094] When the maximum stress rate is far less than 1, it means that the safe operating threshold of the device is too high and over-design occurs. At this time, the device needs to be reselected to reduce the implementation cost, volume and quality to improve the economic benefits of the design.

[0095] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific process of the failure mode analysis in step 105 is as follows:

[0096] Step 1051, start;

[0097] Step 1052: Model the fault in a system model of the independent power system;

[0098] Step 1053: Set the value of the fault mode group number i to 1, and the value of the system stability number m in the fault mode analysis to 0;

[0099] Step 1054: inject the i-th group of fault modes into the system model of the independent power system;

[0100] Step 1055: System time domain simulation to obtain model output voltage and current simulation data;

[0101] Step 1056: extract steady-state characteristic indicators;

[0102] Step 1057: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1058; otherwise, if the steady-state characteristic index does not meet the design requirements, add 1 to the value of i and return to step 1054.

[0103] Step 1058: Add 1 to the value of the system stability times m in the failure mode analysis;

[0104] Step 1059: extract dynamic characteristic indicators;

[0105] Step 10510: Determine whether the value of i is equal to the maximum value M of the set number of failure modes. If the value of i is equal to M, execute step 10511; otherwise, if the value of i is not equal to M, return to execute step 1054.

[0106] Step 10511, end.

[0107] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific process of determining the independent power system robustness evaluation index in step 201 is as follows:

[0108] Step 211: The probability measure of the robustness of the independent power system is called robustness, and the system robustness γ under the i-th level system dynamic performance index is called i It is expressed as γ i =δ i *w, where i is the system dynamic performance index considered in the robustness calculation, and the value of i is a natural number from 1 to n′, and n′ is the total number of system dynamic performance indexes considered in the robustness calculation; δ i The system dynamic performance index calculated after N statistical analysis Monte Carlo simulations and M failure mode analysis simulations meets the probability level specified in the i-th level and n xi is the number of simulations in which each dynamic performance index x meets the i-th level index in the N-times statistical analysis Monte Carlo simulation, m xi The number of simulations for each dynamic performance index x that meets the i-th level index for M fault mode analysis simulations, x=[x1,x2,…,x λ ], λ is the total number of types of dynamic performance indicators; w is the weight and is the corresponding weight of the dynamic performance index x1 in the independent power system robustness assessment, is the corresponding weight of the dynamic performance index x2 in the independent power system robustness assessment, is the dynamic performance index x λ The corresponding weight in the independent power system robustness assessment;

[0109] Step 2012: According to the formula Calculate the probability level δ of the system dynamic performance index obtained after N statistical analysis Monte Carlo simulations and M failure mode analysis simulations to meet the requirements of level i i ;

[0110] Step 2013: Use empirical method or analytic hierarchy process to assign weight w;

[0111] Among them, the method of giving the weight w by the experience-based method is as follows: the designer directly gives the weight w based on past similar cases or personal experience;

[0112] The process of giving weight w in the hierarchical analysis method is as follows:

[0113] Step 2131. Start: Establish a hierarchical model, dividing the decision problem into a goal layer and an indicator layer. The goal layer contains only one element, which is the predetermined goal and ideal result of the analysis problem. The indicator layer contains the intermediate links involved in achieving the goal, multiple consideration indicators, and multiple criteria.

[0114] Step 20132: Use scales 1 to 9 to construct a comparison matrix A, and compare the values ​​a obtained by comparing different criteria pairwise. i′j′ Fill in the position of row i' and column j' in the matrix, and compare the diagonal of matrix A to make sure all the diagonals are 1;

[0115] Step 20133: Calculate the eigenvalues ​​and eigenvectors of the comparison matrix A;

[0116] Step 20134, according to the formula Calculate the consistency ratio CR, where n″ is the trace of matrix A, λ max is the maximum eigenvalue of the comparison matrix A, RI is the random consistency index;

[0117] Step 20135: Perform a consistency check to determine whether the value of CR is less than 0.1. If the value of CR is less than 0.1, the process ends and the eigenvector of the comparison matrix A obtained in step 20133 is determined as the weight w. Otherwise, if the value of CR is not less than 0.1, the process returns to step 20132.

[0118] Step 2014: According to formula γ i =δ i *w calculates the system robustness γ under the dynamic performance index of the i-th level system i .

[0119] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that: when determining the independent power system robustness evaluation method in step 202, there are two methods, one is to evaluate according to the independent power system robustness evaluation index determined in step 201, and the other is to evaluate according to whether the system is stable; when the independent power system robustness evaluation method is determined to be based on whether the system is stable, assuming that there are n stable times in N statistical analysis Monte Carlo simulations and m stable times in M ​​fault mode analysis simulations, the system robustness calculation formula is simplified as follows:

[0120] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized in that the specific process of performing independent power system parameter analysis based on Pareto analysis in step 301 is as follows:

[0121] Step 3011, start;

[0122] Step 3012: Set the value of i″ to 1;

[0123] Step 3013: Select the i″th model parameter with tolerance setting;

[0124] Step 3014: extract simulation parameter samples and system dynamic performance index results of h Monte Carlo simulations that meet the first-level indicators in the statistical analysis;

[0125] Step 3015: Draw a scatter plot with the parameter value as the horizontal axis and the dynamic performance index as the vertical axis;

[0126] Step 3016: Perform least squares straight line fitting on the scatter plot to obtain the slope of the fitted straight line;

[0127] Step 3017: Calculate statistical sensitivity and correlation coefficient;

[0128] Step 3018: Determine whether the value of i″ is equal to h. If so, execute step 3019. Otherwise, if the value of i″ is not equal to h, add 1 to the value of i″ and return to step 3013.

[0129] Step 3019, end.

[0130] The above-mentioned independent power system robustness analysis and optimization method applicable to MBSE is characterized by: in step 3016, the least squares straight line fitting is performed on the scatter plot, and the straight line expression obtained by fitting is m′=β0+β1p, where m′ represents the system response dynamic performance index, p represents the parameter subjected to disturbance, β0 is the constant term of the fitted straight line, and β1 is the slope of the fitted straight line;

[0131] The formula for calculating the statistical sensitivity in step 3017 is: Among them, P sens is the statistical sensitivity, Σp / Σm′ is the normalization factor, N min The minimum number of simulations required for Monte Carlo simulation;

[0132] The formula for calculating the correlation coefficient in step 3017 is: Among them, P Correlation is the correlation coefficient, ranging from 0 to 1;

[0133] The method for determining parameters for the robust optimization design of the independent power system described in step 302 is: based on the scatter plot drawn in step 3015, find the Pareto front, weigh the feasible solutions in the Pareto front, and select a suitable value through compromise.

[0134] Compared with the prior art, the present invention has the following advantages:

[0135] 1. The present invention can quantitatively evaluate the robustness of different designs of independent power systems, including different protection strategies, different control logics, different design parameters, etc., so as to make candidate design decisions.

[0136] 2. The present invention can optimize the system performance of a given design and enhance the robustness of an independent power system.

[0137] 3. This invention can improve system development efficiency and help reduce R&D costs. It can comprehensively obtain information affecting system performance before the actual experimental platform is built, and provides a robust design optimization method for the system. This reduces the number of iterations of hardware experimental prototypes, thereby optimizing the traditional independent power system development process. This provides an effective solution for optimizing system performance, improving system development efficiency, and reducing development costs.

[0138] 4. The present invention combines MBSE to optimize the design iteration strategy of the independent power system.

[0139] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0140] Figure 1 It is a flowchart of the method of the present invention.

[0141] Figure 2 A flow chart of the independent power system rating design applicable to MBSE according to the present invention;

[0142] Figure 3 This is a flow chart of the sensitivity analysis of an independent power system applicable to MBSE according to the present invention;

[0143] Figure 4 is a flow chart of independent power system statistical analysis applicable to MBSE according to the present invention;

[0144] Figure 5 This is a flow chart of independent power system stress analysis applicable to MBSE according to the present invention;

[0145] Figure 6 This is a flow chart of the independent power system failure mode analysis applicable to MBSE according to the present invention;

[0146] Figure 7 A schematic diagram of a diode fault simulation according to the present invention;

[0147] Figure 8 A schematic diagram of a winding phase-to-phase short circuit fault modeling according to the present invention;

[0148] Figure 9 The flowchart of the present invention using the analytic hierarchy process to assign weights;

[0149] Figure 10 A flowchart for implementing Pareto analysis in the present invention;

[0150] Figure 11A A schematic diagram of parameters that have a significant impact on system performance in the Pareto analysis results of the present invention;

[0151] Figure 11B A schematic diagram of parameters with minimal impact on system performance in the Pareto analysis results of the present invention;

[0152] Figure 12 Schematic diagram of the Pareto front of the present invention. DETAILED DESCRIPTION

[0153] like Figure 1 As shown, the independent power system robustness analysis and optimization method applicable to MBSE of the present invention is characterized in that the method comprises the following steps:

[0154] Step 1: Establish a simulation-based independent power system robustness analysis process, including:

[0155] Step 101, rated design: Build a system model of the independent power system, perform time-domain simulation on the model of the independent power system under rated conditions, obtain simulation data of voltage and current, and use the simulation data of voltage and current to analyze two levels of indicators: steady-state characteristic indicators and dynamic characteristic indicators to ensure that the independent power system performs according to the expected performance specifications under rated conditions;

[0156] The purpose of performing rated design is to use the results of rated design to establish response targets / reference benchmarks for the remaining four analyses in the robustness analysis process. During rated design, as long as the output voltage value of the model can fall within the stable voltage error band (e.g., ±5%) within a certain period of time, the current system stability is considered to be insufficient for engineering practice.

[0157] The steady-state characteristic index in step 101 is defined as a primary index, which is used to reflect the steady-state function of the independent power system and indicates that the current system is in a stable operating state. The dynamic characteristic index in step 101 is defined as a secondary index, which is used to reflect the dynamic performance of the independent power system and measure the dynamic performance of the system when it is subjected to different disturbances, so as to select the best system design under the premise of ensuring that the steady-state performance meets the requirements.

[0158] The specific process of the rated design in step 101 is as follows:

[0159] Step 1011: Start;

[0160] Step 1012: Build a system model of the target independent power system;

[0161] Step 1013: setting the model parameters to a rated state;

[0162] Step 1014: setting the model simulation working condition to the rated state;

[0163] Step 1015: System time domain simulation to obtain model output voltage and current simulation data;

[0164] Step 1016: extract steady-state characteristic indicators;

[0165] Step 1017: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1018; otherwise, if the steady-state characteristic index does not meet the design requirements, return to step 1013.

[0166] Step 1018: extract dynamic characteristic indicators;

[0167] Step 1019: Determine whether the dynamic characteristic index meets the design requirements. If the dynamic characteristic index meets the design requirements, execute step 10110; otherwise, if the dynamic characteristic index does not meet the design requirements, return to step 1013.

[0168] Step 10110, end.

[0169] The flow chart of the rated design described in step 101 is as follows Figure 2 shown.

[0170] In specific implementation, when extracting the first-level and second-level indicators, they shall be extracted according to the relevant national standards involved in the specific independent power system studied by the designers, the provisions of the project technical agreement, and the research objectives of the subject;

[0171] For example, for an independent power system in a ground power grid, the first-level indicators that can be selected include bus voltage, grid frequency, phase, active power, etc.; for the second-level indicators that can be selected include grid frequency overshoot, grid frequency adjustment time, bus voltage drop, etc.

[0172] For example, the aircraft power supply system is a type of independent power system. For the selection of first-level indicators, refer to the provisions of GJB181B, and select power supply quality parameters such as steady-state voltage, distortion coefficient, distortion spectrum, and pulsation amplitude as steady-state characteristic indicators. The specific limits of the above four steady-state characteristic indicators refer to the provisions in GJB181B, and the calculation of the above four steady-state characteristic indicators refers to the provisions in GJB5189; for the selection of second-level indicators, voltage and current overshoot, peak time, adjustment time and other quantities are selected as dynamic characteristic indicators. The calculation of the above three dynamic characteristic indicators refers to the classic definitions in automatic control theory.

[0173] The results of the rated design in the present invention are measured by two-level indicators, which can better determine the response targets / reference benchmarks for the following four analyses in the robustness analysis process.

[0174] Step 102, sensitivity analysis: applying small positive and negative disturbances to each parameter of the system model according to the set parameter disturbance ratio d%, performing system simulation and recording the system response dynamic characteristic indicators, calculating the sensitivity per unit value of different parameters relative to different system response dynamic characteristic indicators, until the sensitivity analysis of all parameters in the system model is completed;

[0175] The purpose of sensitivity analysis is to help users quickly identify which parameters in an individual power system are affecting system performance, establish engineering priorities, and manage the resulting changes to minimize any potential degradation in performance.

[0176] The specific process of the sensitivity analysis in step 102 is as follows:

[0177] Step 1021: Start;

[0178] Step 1022: Set the system model parameter disturbance ratio d%.

[0179] Step 1023: Select the next system model parameter;

[0180] Step 1024: applying small positive disturbances and negative disturbances to the parameters respectively;

[0181] Step 1025: Perform system simulation to obtain the model output voltage and current response;

[0182] Step 1026: extract dynamic performance indicators;

[0183] Step 1027: Calculate the per-unit positive disturbance sensitivity and the per-unit negative disturbance sensitivity of the parameter relative to different dynamic performance indicators respectively;

[0184] The per-unit value of the positive disturbance sensitivity is the per-unit value of the sensitivity obtained under positive disturbance, and the per-unit value of the negative disturbance sensitivity is the per-unit value of the sensitivity obtained under negative disturbance;

[0185] Step 1028: Determine whether the per-unit values ​​of the positive disturbance sensitivity and the per-unit values ​​of the negative disturbance sensitivity of all parameters have been calculated. If the calculation is completed, execute step 1029; otherwise, if the calculation is not completed, return to step 1023.

[0186] Step 1029, end.

[0187] The flowchart of the sensitivity analysis in step 102 is as follows: Figure 3 shown.

[0188] The parameter disturbance ratio d% in step 1022 is set to be less than or equal to 5%;

[0189] The calculation formulas for the per-unit value of the positive disturbance sensitivity and the per-unit value of the negative disturbance sensitivity in step 1027 are: Where Sensitivity′ is the per-unit value of the positive disturbance sensitivity or the per-unit value of the negative disturbance sensitivity, m now is the dynamic performance index value extracted from the sensitivity analysis simulation (i.e., the dynamic performance index value extracted in step 1026), m rated is the dynamic performance index value extracted in the rated design (ie, the secondary index value extracted in step 1018).

[0190] For an independent power system, it is generally believed that a disturbance ratio not higher than 5% is a small disturbance. When performing sensitivity analysis, the disturbance ratio d% of the parameter should not be higher than 5%. Therefore, the value of the disturbance ratio d% in the present invention is less than or equal to 5%.

[0191] In specific implementation, whether it is the positive disturbance sensitivity per unit value or the negative disturbance sensitivity per unit value, the values ​​can be positive or negative; a positive value means that when the parameter undergoes positive / negative disturbance, the dynamic performance index will be larger than the system at the rated design; a negative value means that when the parameter undergoes positive / negative disturbance, the dynamic performance index will be smaller than the system at the rated design; the larger the absolute value of the sensitivity per unit value, the greater the influence of this parameter on the index, and a sensitivity of 0 means no influence.

[0192] By setting a negative perturbation, the present invention can eliminate the situation where positive parameter changes have little effect on the system's dynamic response performance indicators, but negative parameter changes have a significant impact on the system's dynamic response performance indicators. For example, before the knee point c of the magnetization curve of a ferromagnetic material, increasing the external magnetic field strength H will cause the ferromagnetic material's magnetic flux density B to rise rapidly. However, if the external magnetic field strength H is increased after the knee point c, the ferromagnetic material's magnetic flux density B will rise very slowly.

[0193] When performing sensitivity analysis, the present invention applies the same degree of positive disturbance and negative disturbance to each parameter, so as to obtain whether each parameter will have different effects on different dynamic performance indicators when it changes near the operating point, and the degree of the influence.

[0194] Step 103, statistical analysis: simulate noise factors and perform Monte Carlo simulation (MC simulation) based on Latin Hypercube Sampling (LHS) on the independent power system model to obtain information on two levels of indicators, namely steady-state characteristic indicators and dynamic characteristic indicators, for the established system model parameters under uncertain noise factors;

[0195] The purpose of statistical analysis is to: Due to the designer's limited understanding of system parameters, fluctuations in component parameters caused by the manufacturing process, and changes in boundary conditions, the model input parameters are actually uncertain. This is a noise factor that must be considered in system robustness analysis. Statistical analysis performs simulations with uncertain parameters, which also represents the system performance under small perturbations.

[0196] The specific process of the statistical analysis in step 103 is as follows:

[0197] Step 1031: Start;

[0198] Step 1032: Create a list of key system model parameters to be observed;

[0199] The method for creating the list of key parameters of the system model to be observed in step 1032 adopts one of the following three methods:

[0200] Method A1: Use the results of the sensitivity analysis to select the factors that most affect the change in the performance metric and create a list of key system model parameters to be observed.

[0201] Method A2: Use industry knowledge and rules of thumb to analyze the type of system being analyzed and create a list of key system model parameters to observe;

[0202] Method A3: Rely on the designer's experience and create a list of key parameters of the system model to be observed based on previous designs of similar systems.

[0203] Step 1033: specify distribution and tolerance range for each key parameter;

[0204] When specifying a distribution for each key parameter in step 1033, the distribution includes normal distribution, average distribution, exponential distribution, and Γ-distribution. In specific implementation, the designer refers to the device parameter manual, targeted experimental test analysis results, and the knowledge and experience of the engineering staff to specify the distribution of each key parameter as one of the above distributions.

[0205] The method for specifying the tolerance range for each key parameter in step 1033 includes the following methods:

[0206] Method B1: For actual electronic component parameters, query the tolerance range of the parameters from the device data sheet, and estimate the tolerance range based on the system operating temperature and component temperature drift;

[0207] Method B2: The tolerance range of mechanical structure parameters is determined according to the manufacturing tolerance range;

[0208] Method B3: For parameters that need to be measured by specific experiments, the tolerance range is determined based on the measurement error of the instrument;

[0209] Method B4: For parameters obtained by fitting measured data, the fitting error is determined as the parameter tolerance range, or the tolerance range is estimated in combination with the instrument measurement error;

[0210] In specific implementation, designers use one of the above four methods to specify the tolerance range for each key parameter;

[0211] Step 1034: Given the estimated accuracy and confidence level, calculate the minimum number of simulations N required for Monte Carlo simulation. min ;

[0212] Given the estimated accuracy and confidence level in step 1034, the minimum number of simulations N required for Monte Carlo simulation is calculated. min When , the calculation formula used is: Where δ is the confidence level and δ∈(0,1), ε is the estimation accuracy and ε∈(0,1);

[0213] Step 1035: Perform Latin Hypercube Sampling (LHS) on the input parameters within the tolerance range determined in step 1033 to generate N min Set parameter combination samples; take the initial value of parameter group number i as 1, and the initial value of system stability times n in statistical analysis as 0;

[0214] When performing Latin Hypercube Sampling (LHS) on the input parameters within the tolerance range determined in step 1033 in step 1035, assuming that there are n′ input uncertain parameters in the model, N min Monte Carlo simulations require N extractions in the n′-dimensional vector space. min The specific process of Latin hypercube sampling (LHS) is as follows:

[0215] Step 10351: Divide each dimensional vector space into N non-overlapping min intervals, so that each interval has the same probability;

[0216] In specific implementation, the interval is usually divided into equal parts;

[0217] Step 10352: randomly select a point in each interval in each dimensional vector space;

[0218] Step 10353: Randomly extract the points selected in step 10352 from the n′-dimensional vector space to form N min Set vector, the N min The group vector is determined as N for Monte Carlo simulation min Group parameter combination samples.

[0219] Step 1036: inject the i-th group of parameters into the system model;

[0220] Step 1037: System time domain simulation to obtain model output voltage and current simulation data;

[0221] Step 1038: extract steady-state characteristic indicators;

[0222] Step 1039: Determine whether the steady-state characteristic index meets the design requirements. If so, execute step 10310. Otherwise, if not, increment the value of i by 1, and return to step 1036.

[0223] Step 10310: Add 1 to the value of the system stability times n in the statistical analysis;

[0224] Step 10311: extract dynamic characteristic indicators;

[0225] Step 10312: Determine whether the value of i is equal to N. min , when the value of i is equal to N min , execute step 10313; otherwise, when the value of i is not equal to N min When , add 1 to the value of i and return to step 1036;

[0226] Step 10313, end.

[0227] The flowchart of the statistical analysis in step 103 is as follows Figure 4 shown.

[0228] Step 104, stress analysis: searching for voltage and current data of each component from simulation results that meet the steady-state characteristic index requirements in the statistical analysis, obtaining the maximum stress rate, and performing stress analysis based on the maximum stress rate;

[0229] The specific process of the stress analysis in step 104 is as follows:

[0230] Step 1041: Start;

[0231] Step 1042: Set the value of i to 1;

[0232] Step 1043: Statistically analyze the i-th Monte Carlo simulation (MC simulation);

[0233] Step 1044: Obtain system model Monte Carlo simulation results and device voltage and current data;

[0234] Step 1045: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1046; otherwise, if the steady-state characteristic index does not meet the design requirements, add 1 to the value of i and return to step 1043.

[0235] Step 1046: Calculate the stress rate based on the device voltage and current data;

[0236] When calculating the stress rate based on the device voltage and current data in step 1046, the calculation formula used to calculate the stress rate StressRatio(i) during the i-th Monte Carlo simulation is: Among them, max(|E i |) is the maximum value of the device voltage and current in the i-th Monte Carlo simulation, rated(|E i |) is the operating threshold of the device voltage and current set by the designer in the i-th Monte Carlo simulation, U i is the voltage value of the device in the i-th Monte Carlo simulation, I i is the current value of the device in the i-th Monte Carlo simulation.

[0237] So, in fact, StressRatio(i) represents the ratio at which the voltage and current exceed the safe operating threshold.

[0238] Step 1047: Determine whether the value of i is equal to N min , when the value of i is equal to N min , execute step 1048; otherwise, when the value of i is not equal to N min When , add 1 to the value of i and return to step 1043;

[0239] Step 1048: From N min Select the maximum stress rate from the calculated stress rates;

[0240] Step 1049: Perform stress analysis based on the maximum stress rate;

[0241] The specific method for performing stress analysis based on the maximum stress rate in step 1049 is:

[0242] When the maximum stress rate is greater than 1, it means that the voltage and current stresses of the device have exceeded the safe operating limits specified by the designer during stress analysis. The operating threshold of the device needs to be reconsidered. This helps designers discover situations where circuit performance does not meet design goals and guides designers in device selection.

[0243] When the maximum stress rate is ≤1, it means that the device has not reached its stress limit value in the stress analysis, and the device selection is reasonable;

[0244] When the maximum stress rate is far less than 1, it means that the safe operating threshold of the device is too high and over-design occurs. At this time, the device needs to be reselected to reduce the implementation cost, volume and quality to improve the economic benefits of the design.

[0245] Step 10410, end.

[0246] The flowchart of the stress analysis in step 104 is as follows: Figure 5 shown.

[0247] Step 105, Fault Mode Analysis: Model the fault to be analyzed in the system model of the independent power system. By setting components to fail in various ways and at specific times, study how single component failures or combined failure modes affect system performance, and obtain analysis results of two-level indicators of the system under the failure modes.

[0248] That is, before starting the fault mode simulation, designers need to model the faults to be analyzed in the system model of the independent power system. The fault modeling method should be determined according to the designer's analysis requirements. During the fault mode analysis process, by setting components to fail in various ways at specific times, the impact of single or combined fault modes on system performance is studied, and the analysis results of the two-level indicators of the system under the fault mode are obtained.

[0249] The purpose of failure mode analysis is to study how single or combined failure modes of components in an independent power system affect system performance, and then determine the severity of the impact. This provides a basis for users to discover potential problems, modify designs, and formulate protection strategies, thereby greatly improving the robustness of the system under large disturbances.

[0250] The specific process of the failure mode analysis in step 105 is as follows:

[0251] Step 1051, start;

[0252] Step 1052: Model the fault in a system model of the independent power system;

[0253] Step 1053: Set the value of the fault mode group number i to 1, and the value of the system stability number m in the fault mode analysis to 0;

[0254] Step 1054: inject the i-th group of fault modes into the system model of the independent power system;

[0255] Step 1055: System time domain simulation to obtain model output voltage and current simulation data;

[0256] Step 1056: extract steady-state characteristic indicators;

[0257] Step 1057: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1058; otherwise, if the steady-state characteristic index does not meet the design requirements, add 1 to the value of i and return to step 1054.

[0258] Step 1058: Add 1 to the value of the system stability times m in the failure mode analysis;

[0259] Step 1059: extract dynamic characteristic indicators;

[0260] Step 10510: Determine whether the value of i is equal to the maximum value M of the set number of failure modes. If the value of i is equal to M, execute step 10511; otherwise, if the value of i is not equal to M, return to execute step 1054.

[0261] Step 10511, end.

[0262] The flowchart of the failure mode analysis in step 105 is as follows: Figure 6 shown.

[0263] Example of parameter fault modeling: Taking diode fault as an example, a diode fault simulation model is established. The fault simulation is performed by directly modifying the fault component parameters. When the diode is normal, R = 0.5827Ω. When the diode is open, R = 0.5297Ω is converted to R = 1×10 12 Ω, when the diode is short-circuited, R = 0.5297Ω is transformed into R = 0Ω, such as Figure 7 As shown;

[0264] Example of structural fault modeling: Taking winding phase short circuit / ground short circuit as an example, there are two implementation methods: Figure 8 The first type is to connect a resistor with a minimum value in series, such as between the A-phase and B-phase windings. A resistor R is connected in series between the groups. When no interphase short circuit fault occurs, the resistance R value is extremely large. When an interphase short circuit fault occurs, R is changed to 0Ω. The second type is to connect a switch in series, such as between the B-phase and C-phase windings. A switch is connected in series between the groups. When no interphase short circuit fault occurs, the switch is disconnected. When an interphase short circuit fault occurs, the switch is closed.

[0265] Step 2: Conduct an independent power system robustness assessment based on robustness, including:

[0266] Step 201: Determine an independent power system robustness assessment index;

[0267] The specific process of determining the independent power system robustness evaluation index in step 201 is as follows:

[0268] Step 211: The probability measure of the robustness of the independent power system is called robustness, and the system robustness γ under the i-th level system dynamic performance index is called i It is expressed as γ i =δ i*w, where i is the system dynamic performance index considered in the robustness calculation, and the value of i is a natural number from 1 to n′, and n′ is the total number of system dynamic performance indexes considered in the robustness calculation; δ i The system dynamic performance index calculated after N statistical analysis Monte Carlo simulations and M failure mode analysis simulations meets the probability level specified in the i-th level and n xi is the number of simulations in which each dynamic performance index x meets the i-th level index in the N-times statistical analysis Monte Carlo simulation, m xi The number of simulations for each dynamic performance index x that meets the i-th level index for M fault mode analysis simulations, x=[x1,x2,…,x λ ], λ is the total number of types of dynamic performance indicators; w is the weight and is the corresponding weight of the dynamic performance index x1 in the independent power system robustness assessment, is the corresponding weight of the dynamic performance index x2 in the independent power system robustness assessment, is the dynamic performance index x λ The corresponding weight in the independent power system robustness assessment;

[0269] Step 2012: According to the formula Calculate the probability level δ of the system dynamic performance index obtained after N statistical analysis Monte Carlo simulations and M failure mode analysis simulations to meet the requirements of level i i ;

[0270] Step 2013: Use empirical method or analytic hierarchy process to assign weight w;

[0271] Among them, the method of giving the weight w by the experience-based method is as follows: the designer directly gives the weight w based on past similar cases or personal experience;

[0272] Among them, such as Figure 9 As shown in Figure 2, the process of giving weight w by AHP is as follows:

[0273] Step 2131. Start: Establish a hierarchical model, dividing the decision problem into a target layer and an indicator layer. The target layer contains only one element, which is the predetermined goal and ideal result of the analysis problem, such as selecting a suitable controller. The indicator layer contains the intermediate links involved in achieving the goal, multiple consideration indicators, and multiple criteria. In this embodiment, these are dynamic performance indicators: overshoot, peak time, and settling time.

[0274] Step 20132: Use scales 1 to 9 to construct a comparison matrix A, and compare the values ​​a obtained by comparing different criteria pairwise. i′j′Fill in the position of row i' and column j' in the matrix, and compare the diagonal of matrix A to make sure all the diagonals are 1;

[0275] Assign a total weight of 1 to the target layer, and then assign the total weight to different indicators in the indicator layer. The weight of the corresponding indicator represents the importance of the factor in the system robustness assessment process. Since it is difficult to directly assign weights to factors of different natures, all factors are not compared together, but are compared with each other. Therefore, in the specific implementation, the comparison matrix A is constructed using scales 1 to 9 as shown in Table 1 to minimize the difficulty of comparing factors of different natures and improve the accuracy of weight assignment.

[0276] Table 1 Meaning of scale

[0277]

[0278] Step 20133: Calculate the eigenvalues ​​and eigenvectors of the comparison matrix A;

[0279] In specific implementation, the maximum eigenvalue λ of the comparison matrix A is calculated max , and λ max The corresponding normalized eigenvector w=[w1,...,w n ] T ;

[0280] Step 20134, according to the formula Calculate the consistency ratio CR, where n″ is the trace of matrix A, λ max is the maximum eigenvalue of the comparison matrix A, RI is the random consistency index;

[0281] In specific implementation, the value of RI is obtained by looking up Table 2;

[0282] Table 2 Correspondence between RI values ​​and n

[0283]

[0284] Step 20135: Perform a consistency check to determine whether the value of CR is less than 0.1. If the value of CR is less than 0.1, the process ends and the eigenvector of the comparison matrix A obtained in step 20133 is determined as the weight w. Otherwise, if the value of CR is not less than 0.1, the process returns to step 20132.

[0285] Step 2014: According to formula γ i =δ i *w calculates the system robustness γ under the dynamic performance index of the i-th level system i .

[0286] Step 202: Determine an independent power system robustness assessment method;

[0287] When determining the independent power system robustness assessment method in step 202, there are two methods: one is to assess according to the independent power system robustness assessment index determined in step 201, and the other is to assess according to whether the system is stable. When the independent power system robustness assessment method is determined to assess according to whether the system is stable, assuming that there are n stable times in N statistical analysis Monte Carlo simulations and m stable times in M ​​fault mode analysis simulations, the robustness calculation formula of the system is simplified as follows:

[0288] Step 203: Evaluate the robustness of the independent power system according to the independent power system robustness evaluation method determined in step 202;

[0289] Step 3: Figure 10 As shown in Figure 2, the robust design optimization of independent power systems based on Pareto analysis includes:

[0290] Step 301: Perform independent power system parameter analysis based on Pareto analysis;

[0291] The specific process of performing independent power system parameter analysis based on Pareto analysis in step 301 is as follows:

[0292] Step 3011, start;

[0293] Step 3012: Set the value of i″ to 1;

[0294] Step 3013: Select the i″th model parameter with tolerance setting;

[0295] Step 3014: extract simulation parameter samples and system dynamic performance index results of h Monte Carlo simulations that meet the first-level indicators in the statistical analysis;

[0296] Step 3015: Draw a scatter plot with the parameter value as the horizontal axis and the dynamic performance index as the vertical axis;

[0297] Step 3016: Perform least squares straight line fitting on the scatter plot to obtain the slope of the fitted straight line;

[0298] In step 3016, the scatter plot is fitted with a least squares straight line, and the fitted straight line expression is m′=β0+β1p, where m′ represents the system response dynamic performance index, p represents the parameter subjected to disturbance, β0 is the constant term of the fitted straight line, and β1 is the slope of the fitted straight line;

[0299] Step 3017: Calculate statistical sensitivity and correlation coefficient;

[0300] The formula for calculating the statistical sensitivity in step 3017 is: Among them, P sens is the statistical sensitivity, Σp / Σm′ is the normalization factor, N min is the minimum number of simulations required for Monte Carlo simulation; the larger the slope β1, the greater the P sens The higher; when P sens When it is greater than 0, it means that the corresponding dynamic performance index will increase as the parameter increases and decrease as the parameter decreases. sens When it is less than 0, it means that the corresponding dynamic performance index will decrease as the parameter increases, and increase as the parameter decreases;

[0301] Statistical sensitivity P sens The connotation of the sensitivity per unit value is similar to that obtained in the sensitivity analysis, but slightly different in that the Pareto analysis includes the effects of parameter interactions or nonlinearity. This is because in the statistical analysis MC simulation process, all parameters are randomly selected as long as they are within their tolerance range, while the sensitivity analysis does not include the effects of parameter interactions.

[0302] The formula for calculating the correlation coefficient in step 3017 is: Among them, P Correlation is the correlation coefficient, ranging from 0 to 1; P Correlation The smaller the value, the less relevant the parameter is to the corresponding dynamic performance index, and vice versa.

[0303] Step 3018: Determine whether the value of i″ is equal to h. If so, execute step 3019. Otherwise, if the value of i″ is not equal to h, add 1 to the value of i″ and return to step 3013.

[0304] Step 3019, end.

[0305] Step 302: Perform parameter determination for robust optimization design of the independent power system.

[0306] The method for determining parameters for the robust optimization design of the independent power system described in step 302 is: based on the scatter plot drawn in step 3015, find the Pareto front, weigh the feasible solutions in the Pareto front, and select a suitable value through compromise.

[0307] For example, in Figure 12 In the Pareto frontier of overshoot and capacitance values ​​shown, in order to obtain the lowest possible overshoot, the capacitance value will increase, resulting in an increase in the volume and mass of the capacitor.

[0308] Example of Pareto analysis results: Figure 11B Parameter 2 shows a small effect on the systematic variation, as its correlation coefficient and statistical sensitivity are low, and according to this particular metric, Figure 11AThe change of parameter 1 has a much greater impact on system performance. Increasing the two parameters in Figure 11 will reduce the corresponding system dynamic performance indicators. When increasing the same proportion, the corresponding system dynamic performance indicators decrease more with parameter 1.

[0309] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A robustness analysis and optimization method for an independent power system suitable for MBSE, characterized by: The method comprises the following steps: Step 1: Establish a simulation-based independent power system robustness analysis process, including: Step 101, rated design: Build a system model of the independent power system, perform time-domain simulation on the model of the independent power system under rated conditions, obtain simulation data of voltage and current, and use the simulation data of voltage and current to analyze two levels of indicators: steady-state characteristic indicators and dynamic characteristic indicators to ensure that the independent power system performs according to the expected performance specifications under rated conditions; Step 102, sensitivity analysis: applying small positive and negative disturbances to each parameter of the system model according to the set parameter disturbance ratio d%, performing system simulation and recording the system response dynamic characteristic indicators, calculating the sensitivity per unit value of different parameters relative to different system response dynamic characteristic indicators, until the sensitivity analysis of all parameters in the system model is completed; Step 103, statistical analysis: simulate noise factors and perform Monte Carlo simulation based on Latin hypercube sampling on the independent power system model to obtain information on two levels of indicators, namely steady-state characteristic indicators and dynamic characteristic indicators, for the established system model parameters under uncertain noise factors; Step 104, stress analysis: searching for voltage and current data of each component from simulation results that meet the steady-state characteristic index requirements in the statistical analysis, obtaining the maximum stress rate, and performing stress analysis based on the maximum stress rate; Step 105, Fault Mode Analysis: Model the fault to be analyzed in the system model of the independent power system. By setting components to fail in various ways and at specific times, study how single component failures or combined failure modes affect system performance, and obtain analysis results of two-level indicators of the system under the failure modes. Step 2: Conduct an independent power system robustness assessment based on robustness, including: Step 201: Determine an independent power system robustness assessment index; Step 202: Determine an independent power system robustness assessment method; Step 203: Evaluate the robustness of the independent power system according to the independent power system robustness evaluation method determined in step 202; Step 3: Perform robust design optimization of the independent power system based on Pareto analysis, including: Step 301: Perform independent power system parameter analysis based on Pareto analysis; Step 302: Perform parameter determination for robust optimization design of the independent power system.

2. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: The steady-state characteristic indicators described in step 101 and step 103 are defined as primary indicators, which are used to reflect the steady-state function of the independent power system and indicate that the current system is in a stable operating state; the dynamic characteristic indicators described in step 101 and step 103 are defined as secondary indicators, which are used to reflect the dynamic performance of the independent power system and measure the dynamic performance of the system when it is subjected to different disturbances, so as to achieve the best system design while ensuring that the steady-state performance meets the requirements.

3. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: The specific process of the rated design in step 101 is as follows: Step 1011: Start; Step 1012: Build a system model of the target independent power system; Step 1013: setting the model parameters to a rated state; Step 1014: setting the model simulation working condition to the rated state; Step 1015: System time domain simulation to obtain model output voltage and current simulation data; Step 1016: extract steady-state characteristic indicators; Step 1017: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1018; otherwise, if the steady-state characteristic index does not meet the design requirements, return to step 1013. Step 1018: extract dynamic characteristic indicators; Step 1019: Determine whether the dynamic characteristic index meets the design requirements. If the dynamic characteristic index meets the design requirements, execute step 10110; otherwise, if the dynamic characteristic index does not meet the design requirements, return to step 1013. Step 10110, end.

4. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: The specific process of the sensitivity analysis in step 102 is as follows: Step 1021: Start; Step 1022: Set the system model parameter disturbance ratio d%. Step 1023: Select the next system model parameter; Step 1024: applying small positive disturbances and negative disturbances to the parameters respectively; Step 1025: Perform system simulation to obtain the model output voltage and current response; Step 1026: extract dynamic performance indicators; Step 1027: Calculate the per-unit positive disturbance sensitivity and the per-unit negative disturbance sensitivity of the parameter relative to different dynamic performance indicators respectively; Step 1028: Determine whether the per-unit values ​​of the positive disturbance sensitivity and the per-unit values ​​of the negative disturbance sensitivity of all parameters have been calculated. If the calculation is completed, execute step 1029; otherwise, if the calculation is not completed, return to step 1023. Step 1029, end.

5. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 4 is characterized in that: The parameter disturbance ratio d% in step 1022 is set to be less than or equal to 5%; The calculation formulas for the per-unit value of the positive disturbance sensitivity and the per-unit value of the negative disturbance sensitivity in step 1027 are: Where Sensitivity′ is the per-unit value of the positive disturbance sensitivity or the per-unit value of the negative disturbance sensitivity, m now is the dynamic performance index value extracted from the sensitivity analysis simulation, m rated is the dynamic performance index value extracted from the rated design.

6. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: The specific process of the statistical analysis in step 103 is as follows: Step 1031: Start; Step 1032: Create a list of key system model parameters to be observed; Step 1033: specify distribution and tolerance range for each key parameter; Step 1034: Given the estimated accuracy and confidence level, calculate the minimum number of simulations N required for Monte Carlo simulation. min ; Step 1035: Perform Latin hypercube sampling on the input parameters within the tolerance range determined in step 1033 to generate N that meet the distribution min Set parameter combination samples; take the initial value of parameter group number i as 1, and the initial value of system stability times n in statistical analysis as 0; Step 1036: inject the i-th group of parameters into the system model; Step 1037: System time domain simulation to obtain model output voltage and current simulation data; Step 1038: extract steady-state characteristic indicators; Step 1039: Determine whether the steady-state characteristic index meets the design requirements. If so, execute step 10310. Otherwise, if not, increment the value of i by 1, and return to step 1036. Step 10310: Add 1 to the value of the system stability times n in the statistical analysis; Step 10311: extract dynamic characteristic indicators; Step 10312: Determine whether the value of i is equal to N. min , when the value of i is equal to N min When , execute step 10313; Otherwise, when the value of i is not equal to N min When , add 1 to the value of i and return to step 1036; Step 10313, end.

7. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 6 is characterized by: The method for creating the list of key parameters of the system model to be observed in step 1032 is one of the following three methods: Method A1: Use the results of the sensitivity analysis to select the factors that most affect the change in the performance metric and create a list of key system model parameters to be observed. Method A2: Use industry knowledge and rules of thumb to analyze the type of system being analyzed and create a list of key system model parameters to observe; Method A3: Rely on the designer's experience and create a list of key parameters of the system model to be observed based on previous designs of similar systems. When specifying a distribution for each key parameter in step 1033, the distribution includes normal distribution, average distribution, exponential distribution, and Γ-distribution; The method for specifying the tolerance range for each key parameter in step 1033 includes the following methods: Method B1: For actual electronic component parameters, query the tolerance range of the parameters from the device data sheet, and estimate the tolerance range based on the system operating temperature and component temperature drift; Method B2: The tolerance range of mechanical structure parameters is determined according to the manufacturing tolerance range; Method B3: For parameters that need to be measured by specific experiments, the tolerance range is determined based on the measurement error of the instrument; Method B4: For parameters obtained by fitting measured data, the fitting error is determined as the parameter tolerance range, or the tolerance range is estimated in combination with the instrument measurement error; Given the estimated accuracy and confidence level in step 1034, the minimum number of simulations N required for Monte Carlo simulation is calculated. min When , the calculation formula used is: Where δ is the confidence level and δ∈(0,1), ε is the estimation accuracy and ε∈(0,1); When performing Latin hypercube sampling on the input parameters within the tolerance range determined in step 1033 in step 1035, assuming that there are n′ input uncertain parameters in the model, N min Monte Carlo simulations require N extractions in the n′-dimensional vector space. min The specific process of Latin hypercube sampling is: Step 10351: Divide each dimensional vector space into N non-overlapping min intervals, so that each interval has the same probability; Step 10352: randomly select a point in each interval in each dimensional vector space; Step 10353: Randomly extract the points selected in step 10352 from the n′-dimensional vector space to form N min Set vector, the N min The group vector is determined as N for Monte Carlo simulation min Group parameter combination samples.

8. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized by: The specific process of the stress analysis in step 104 is as follows: Step 1041: Start; Step 1042: Set the value of i to 1; Step 1043: Statistically analyze the i-th Monte Carlo simulation; Step 1044: Obtain system model Monte Carlo simulation results and device voltage and current data; Step 1045: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1046. Otherwise, when the steady-state characteristic index does not meet the design requirements, the value of i is increased by 1, and the process returns to step 1043; Step 1046: Calculate the stress rate based on the device voltage and current data; Step 1047: Determine whether the value of i is equal to N min , when the value of i is equal to N min When , execute step 1048; Otherwise, when the value of i is not equal to N min When , add 1 to the value of i and return to step 1043; Step 1048: From N min Select the maximum stress rate from the calculated stress rates; Step 1049: Perform stress analysis based on the maximum stress rate; Step 10410, end.

9. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 8, characterized in that: When calculating the stress rate based on the device voltage and current data in step 1046, the calculation formula used to calculate the stress rate StressRatio(i) during the i-th Monte Carlo simulation is: E i =[U i I i ], where max(|E i |) is the maximum value of the device voltage and current in the i-th Monte Carlo simulation, rated(|E i |) is the operating threshold of the device voltage and current set by the designer in the i-th Monte Carlo simulation, U i is the voltage value of the device in the i-th Monte Carlo simulation, I i is the current value of the device in the i-th Monte Carlo simulation.

10. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 8 or 9, characterized in that: The specific method for performing stress analysis based on the maximum stress rate in step 1049 is: When the maximum stress rate is greater than 1, it means that the voltage and current stresses of the device have exceeded the safe operating limits specified by the designer during stress analysis. The operating threshold of the device needs to be reconsidered. This helps designers discover situations where circuit performance does not meet design goals and guides designers in device selection. When the maximum stress rate is ≤1, it means that the device has not reached its stress limit value in the stress analysis, and the device selection is reasonable; When the maximum stress rate is far less than 1, it means that the safe operating threshold of the device is too high and over-design occurs. At this time, the device needs to be reselected to reduce the implementation cost, volume and quality to improve the economic benefits of the design.

11. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: The specific process of the failure mode analysis in step 105 is as follows: Step 1051, start; Step 1052: Model the fault in a system model of the independent power system; Step 1053: Set the value of the fault mode group number q to 1, and the value of the system stability number m in the fault mode analysis to 0; Step 1054: inject the qth group of fault modes into the system model of the independent power system; Step 1055: System time domain simulation to obtain model output voltage and current simulation data; Step 1056: extract steady-state characteristic indicators; Step 1057: Determine whether the steady-state characteristic index meets the design requirements. If the steady-state characteristic index meets the design requirements, execute step 1058. Otherwise, when the steady-state characteristic index does not meet the design requirements, the value of q is increased by 1, and the process returns to step 1054; Step 1058: Add 1 to the value of the system stability times m in the failure mode analysis; Step 1059: extract dynamic characteristic indicators; Step 10510: Determine whether the value of q is equal to the maximum value M of the set fault mode number. If the value of q is equal to M, execute step 10511. Otherwise, when the value of q is not equal to M, return to step 1054; Step 10511, end.

12. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: The specific process of determining the independent power system robustness evaluation index in step 201 is as follows: Step 211: The probability measure of the robustness of the independent power system is called robustness, and the system robustness γ under the i-th level system dynamic performance index is called i It is expressed as γ i =δ i *w, where i is the system dynamic performance index considered in the robustness calculation, and the value of i is a natural number from 1 to n′, and n′ is the total number of system dynamic performance indexes considered in the robustness calculation; δ i The system dynamic performance index calculated after N statistical analysis Monte Carlo simulations and M failure mode analysis simulations meets the probability level specified in the i-th level and n xi is the number of simulations in which each dynamic performance index x meets the i-th level index in the N-times statistical analysis Monte Carlo simulation, m xi The number of simulations for each dynamic performance index x that meets the i-th level index for M fault mode analysis simulations, x=[x1,x2,…,x λ ], λ is the total number of types of dynamic performance indicators; w is the weight and is the corresponding weight of the dynamic performance index x1 in the independent power system robustness assessment, is the corresponding weight of the dynamic performance index x2 in the independent power system robustness assessment, is the dynamic performance index x λ The corresponding weight in the independent power system robustness assessment; Step 2012: According to the formula Calculate the probability level δ of the system dynamic performance index obtained after N statistical analysis Monte Carlo simulations and M failure mode analysis simulations to meet the requirements of level i i ; Step 2013: Use empirical method or analytic hierarchy process to assign weight w; Among them, the method of giving the weight w by the experience-based method is as follows: the designer directly gives the weight w based on past similar cases or personal experience; The process of giving weight w in the hierarchical analysis method is as follows: Step 2131. Start: Establish a hierarchical model, dividing the decision problem into a goal layer and an indicator layer. The goal layer contains only one element, which is the predetermined goal and ideal result of the analysis problem. The indicator layer contains the intermediate links involved in achieving the goal, multiple consideration indicators, and multiple criteria. Step 20132: Use scales 1 to 9 to construct a comparison matrix A, and compare the values ​​a obtained by comparing different criteria pairwise. i′j′ Fill in the position of row i' and column j' in the matrix, and compare the diagonal of matrix A to make sure all the diagonals are 1; Step 20133: Calculate the eigenvalues ​​and eigenvectors of the comparison matrix A; Step 20134, according to the formula Calculate the consistency ratio CR, where n″ is the trace of matrix A, λ max is the maximum eigenvalue of the comparison matrix A, RI is the random consistency index; Step 20135: Perform a consistency check to determine whether the value of CR is less than 0.

1. If the value of CR is less than 0.1, the process ends and the eigenvector of the comparison matrix A obtained in step 20133 is determined as the weight w. Otherwise, if the value of CR is not less than 0.1, the process returns to step 20132. Step 2014: According to formula γ i =δ i *w calculates the system robustness γ under the dynamic performance index of the i-th level system i .

13. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: When determining the independent power system robustness assessment method in step 202, there are two methods: one is to assess based on the independent power system robustness assessment index determined in step 201, and the other is to assess based on whether the system is stable; When the independent power system robustness assessment method is determined to be based on whether the system is stable, assuming that there are n stable times in N statistical analysis Monte Carlo simulations and m stable times in M ​​fault mode analysis simulations, the system robustness calculation formula is simplified as follows:

14. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 1 is characterized in that: The specific process of performing independent power system parameter analysis based on Pareto analysis in step 301 is as follows: Step 3011, start; Step 3012: Set the value of i″ to 1; Step 3013: Select the i″th model parameter with tolerance setting; Step 3014: extract simulation parameter samples and system dynamic performance index results of h Monte Carlo simulations that meet the first-level indicators in the statistical analysis; Step 3015: Draw a scatter plot with the parameter value as the horizontal axis and the dynamic performance index as the vertical axis; Step 3016: Perform least squares straight line fitting on the scatter plot to obtain the slope of the fitted straight line; Step 3017: Calculate statistical sensitivity and correlation coefficient; Step 3018: Determine whether the value of i″ is equal to h. If so, execute step 3019. Otherwise, if the value of i″ is not equal to h, add 1 to the value of i″ and return to step 3013. Step 3019, end.

15. The independent power system robustness analysis and optimization method applicable to MBSE according to claim 14, characterized in that: In step 3016, the scatter plot is fitted with a least squares straight line, and the fitted straight line expression is m′=β0+β1p, where m′ represents the system response dynamic performance index, p represents the parameter subjected to disturbance, β0 is the constant term of the fitted straight line, and β1 is the slope of the fitted straight line; The formula for calculating the statistical sensitivity in step 3017 is: Among them, P sens is the statistical sensitivity, Σp / Σm′ is the normalization factor, N min The minimum number of simulations required for Monte Carlo simulation; The formula for calculating the correlation coefficient in step 3017 is: Among them, P Correlation is the correlation coefficient, ranging from 0 to 1; The method for determining parameters for the robust optimization design of the independent power system described in step 302 is: based on the scatter plot drawn in step 3015, find the Pareto front, weigh the feasible solutions in the Pareto front, and select a suitable value through compromise.

Images