技术领域technical field
本发明涉及微电子封装信号完整性技术领域，具体是基于响应曲面法和遗传算法优化PCB微带线结构的方法。The invention relates to the technical field of microelectronic packaging signal integrity, in particular to a method for optimizing a PCB microstrip line structure based on a response surface method and a genetic algorithm.
背景技术Background technique
随着大规模集成电路正在迅速地向高速度、高频率、高密度方向发展，时钟频率可达到几百MHz乃至数GHz，数据率达Gbps以上，电路板上甚至集成数万以上的电子元器件。信号在器件与器件、芯片与芯片之间的传输离不开互连结构，但是随着元器件密集度的增大，互连结构的密度随之变的更为紧凑，尤其是微带线的间距已经达到um级别。在频率日益增高、互连结构密度不断增大互连、结构尺寸也在不断减小的情况下,互连结构所传输的高速脉冲信号在频谱高端对应的波长已与互连结构的尺寸处于同一数量级，信号脉冲在互连线上呈现明显的波动效应，此时互连线已不是简单的连接线而作为具有寄生效应的多导体传输线处理。寄生效应会在传输信号上引起噪声和干扰，使得高速互连结构信号完整性问题变得越来越突出。微带线作为互连结构中关键部分，必须确保电流和信号的正确传输，在高速高频条件下，如果无法保证信号在传输线内的正确传输，即会造成整个系统性能的下降，因此对微带线展开信号完整性问题的分析极为必要。With the rapid development of large-scale integrated circuits in the direction of high speed, high frequency and high density, the clock frequency can reach hundreds of MHz or even several GHz, the data rate can reach more than Gbps, and even tens of thousands of electronic components are integrated on the circuit board. . The transmission of signals between devices and between chips is inseparable from the interconnection structure, but as the density of components increases, the density of the interconnection structure becomes more compact, especially the microstrip line. The spacing has reached the um level. As the frequency increases, the density of the interconnection structure increases, and the size of the structure decreases, the wavelength corresponding to the high-end spectrum of the high-speed pulse signal transmitted by the interconnection structure is already at the same level as the size of the interconnection structure. On the order of magnitude, the signal pulse presents an obvious fluctuating effect on the interconnection line. At this time, the interconnection line is no longer a simple connection line but treated as a multi-conductor transmission line with parasitic effects. Parasitic effects will cause noise and interference on the transmitted signal, making the signal integrity problem of high-speed interconnect structures more and more prominent. As a key part of the interconnect structure, the microstrip line must ensure the correct transmission of current and signals. Under high-speed and high-frequency conditions, if the correct transmission of signals in the transmission line cannot be guaranteed, the performance of the entire system will be reduced. It is extremely necessary to analyze the signal integrity problem of stripline expansion.
由于工程结构的复杂性，结构的功能常常无法直接用结构设计的随机变量做函数表达，因此不能直接运用一次二阶矩方法计算，于是BOX和Wilson提出了响应面法。响应面法也称为回归分析，是数学方法和数理统计结合的产物，是一种用近似的函数关系式表示变量与目标的拟合设计方法。该方法首先利用中心复、Box-Behnken设计、均匀等实验设计、均匀等实验方法建立因素的若干实验组合，分别对各进行获得相应目标值然后选择方法建立因素的若干实验组合，然后选择合适的数学模型对因素与目标结果表示，再运用最小二乘原理求得中未知系数，最后得到变量与结果的拟合函数表达式。RSM能通过较少的实验次数在一定范围内比较精确地逼近因素与目标值之间的函数关系，并用简单表达式展现出来，而且通过对回归模型的选择在一定范围内可以拟复杂响应关系，具有优良鲁棒性能，计算较为简单，为后期参数优化设计带来极大方便。Due to the complexity of the engineering structure, the function of the structure often cannot be directly expressed as a function of the random variable of the structural design, so it cannot be directly calculated by the first-order second-order moment method, so BOX and Wilson proposed the response surface method. Response surface method, also known as regression analysis, is the product of the combination of mathematical methods and mathematical statistics. It is a fitting design method that uses approximate functional relationships to represent variables and targets. This method first uses several experimental combinations of experimental methods such as central complex, Box-Behnken design, uniform experimental design, uniform and other experimental methods to establish factors, and then conducts several experimental combinations to obtain corresponding target values and then selects the method to establish factors, and then selects the appropriate one. The mathematical model expresses the factors and target results, and then uses the least square principle to obtain the unknown coefficients, and finally obtains the fitting function expressions of variables and results. RSM can accurately approximate the functional relationship between the factors and the target value within a certain range through a small number of experiments, and display it with a simple expression, and through the selection of the regression model, it can simulate complex response relationships within a certain range. It has excellent robust performance, and the calculation is relatively simple, which brings great convenience to the later parameter optimization design.
遗传算法是计算数学中的一种全局优化算法，非常适合解决大规模的组合优化问题。电子元件的布局属于组合优化中的旅行商(TSP)问题，近年来已有学者将遗传算法应用到该领域研究中，因此，采用标准遗传算法进行优化可以得到比较好的结果，容易实现优化效果。Genetic algorithm is a global optimization algorithm in computational mathematics, which is very suitable for solving large-scale combinatorial optimization problems. The layout of electronic components belongs to the traveling salesman (TSP) problem in combinatorial optimization. In recent years, some scholars have applied genetic algorithms to research in this field. Therefore, using standard genetic algorithms for optimization can get better results, and it is easy to achieve optimization results. .
发明内容Contents of the invention
本发明的目的在于克服现有技术的不足，而提供一种响应曲面法-遗传算法优化PCB微带线结构的方法，该方法具有优良鲁棒性能，计算较为简单，为后期参数优化设计带来极大方便，优化后的计算结果较为理想。The purpose of the present invention is to overcome the deficiencies in the prior art, and provide a method for optimizing PCB microstrip line structure by response surface method-genetic algorithm. It is very convenient, and the calculation result after optimization is ideal.
实现本发明目的的技术方案是：The technical scheme that realizes the object of the present invention is:
基于响应曲面法和遗传算法优化PCB微带线结构的方法，先利用响应面法设计29组实验组合，根据这29组实验参数，建立相应的29组仿真模型，已利用响应曲面法得到5GHZ下的回波损耗S11与关键因素的函数关系式，在对所获得的函数式进行方差分析，确定了回归方程的有效性；再利用遗传算法对回归方程进行优化，依次执行初始种群生成、交叉、变异和进化逆转操作，获得最有利于微带线信号传输的最优组合，最后通过建立HFSS仿真模型和制作试验样件测量得以验证，具体包括以下步骤：Based on the response surface method and genetic algorithm to optimize the structure of PCB microstrip lines, first use the response surface method to design 29 sets of experimental combinations, according to the 29 sets of experimental parameters, establish the corresponding 29 sets of simulation models, and use the response surface method to obtain 5GHZ. The functional relationship between the return loss S11 and the key factors, the variance analysis of the obtained functional formula is carried out to confirm the validity of the regression equation; then the genetic algorithm is used to optimize the regression equation, and the initial population generation, crossover, The mutation and evolutionary reversal operations are used to obtain the optimal combination that is most beneficial to the signal transmission of the microstrip line. Finally, it is verified by establishing a HFSS simulation model and making a test sample measurement, which specifically includes the following steps:
步骤1：建立HFSS微带线信号完整性分析模型；Step 1: Establish an HFSS microstrip line signal integrity analysis model;
步骤2：获取微带线的回波损耗以及插入损耗；Step 2: Obtain the return loss and insertion loss of the microstrip line;
步骤3：确立影响回波损耗的影响因素；Step 3: Establish the influencing factors affecting the return loss;
步骤4：确立影响因素的参数水平值；Step 4: Establish the parameter level value of the influencing factors;
步骤5：采用BOX-Behnken的中心组合设计模型设计需要的29组实验样本；Step 5: Use BOX-Behnken's central composite design model to design 29 groups of experimental samples;
步骤6：获得影响因素与回波损耗的函数关系式；Step 6: Obtain the functional relationship between the influencing factors and the return loss;
步骤7：对所得函数关系是进行方差分析；Step 7: carry out analysis of variance to the obtained functional relationship;
步骤8：确立所得函数关系式的正确性；Step 8: Establish the correctness of the obtained functional relationship;
步骤9：采用随机方式生成初始种群；Step 9: Generate the initial population randomly;
步骤10：获得当前进化代数gen和最优适应度值；Step 10: Obtain the current evolution algebra gen and optimal fitness value;
步骤11：分别对种群实施交叉操作；Step 11: Carry out the cross operation on the population respectively;
步骤12：分别对种群实施变异操作；Step 12: Implement mutation operations on the populations respectively;
步骤13：分别对种群实施进化逆转；Step 13: Implement evolutionary reversal on the populations respectively;
步骤14：将种群作为整体计算适应度函数值，并采用最优保存策略选择最佳个体；Step 14: Calculate the fitness function value of the population as a whole, and use the optimal preservation strategy to select the best individual;
步骤15：种群更新后重新判断，若gen值小于50且num值大于0，则对种群实施局部灾变。Step 15: Re-judgment after the population is updated, if the gen value is less than 50 and the num value is greater than 0, a local catastrophe will be implemented on the population.
所述步骤1中，模型的尺寸为PCB基板长为15-20mm，宽为5-15mm，高为5-15mm，PCB基板材料为FR4，介电常数为4.4；微带线长为15-20mm，宽为0.1-0.2mm，厚为0.03-0.04mm；参考层厚度0.2-0.4mm。In the step 1, the size of the model is that the length of the PCB substrate is 15-20mm, the width is 5-15mm, and the height is 5-15mm, the material of the PCB substrate is FR4, and the dielectric constant is 4.4; the length of the microstrip line is 15-20mm , the width is 0.1-0.2mm, the thickness is 0.03-0.04mm; the reference layer thickness is 0.2-0.4mm.
所述步骤2中，回波损耗和插入损耗的频率范围为1GHz～5GHz。In the step 2, the frequency range of return loss and insertion loss is 1 GHz-5 GHz.
所述步骤3中，影响因素为基板厚度、微带线宽度、微带线厚度和基板介电常数。In the step 3, the influencing factors are the thickness of the substrate, the width of the microstrip line, the thickness of the microstrip line and the dielectric constant of the substrate.
所述步骤4中，参数水平值的水平数为3，因素数为4。In the step 4, the number of levels of the parameter level value is 3, and the number of factors is 4.
所述步骤5中，是采用BOX-Behnken的中心组合设计模型设计需要的29组实验样本，其中24组为分析因子，5组为零点因子，即参数水平组合相同，用于实验误差估计。In the step 5, the BOX-Behnken central combination design model is used to design 29 groups of experimental samples, of which 24 groups are analysis factors and 5 groups are zero-point factors, that is, the parameter level combinations are the same, and are used for experimental error estimation.
所述步骤6中，分析影响因子与回波损耗的关系在信号频率为5GHZ条件下进行。In the step 6, the analysis of the relationship between the impact factor and the return loss is carried out under the condition that the signal frequency is 5GHZ.
所述步骤9中，种群规模设置为40。In the step 9, the population size is set to 40.
所述步骤10，遗传代数设置为50。In step 10, the genetic algebra is set to 50.
有益效果：该方法通过较少的实验次数在一定范围内比较精确地逼近因素与目标值之间的函数关系，并用简单表达式展现出来，而且通过对回归模型的选择在一定范围内可以拟复杂响应关系，具有优良鲁棒性能，计算较为简单，为后期参数优化设计带来极大方便。Beneficial effects: the method accurately approximates the functional relationship between the factors and the target value within a certain range through a small number of experiments, and displays it with a simple expression, and through the selection of the regression model, it can be simulated within a certain range. The response relationship has excellent robust performance, and the calculation is relatively simple, which brings great convenience to the later parameter optimization design.
附图说明Description of drawings
图1为本发明的基本模型仿真后所得回波损耗图；Fig. 1 is the obtained return loss figure after basic model simulation of the present invention;
图2为本发明的基本模型仿真后所得回波损耗图；Fig. 2 is the obtained return loss figure after basic model simulation of the present invention;
图3为回归方程经过神经网络优化后均值变化图；Fig. 3 is the mean value change diagram after the regression equation is optimized through the neural network;
图4为回归方程经过神经网络优化后最优解的变化图；Fig. 4 is the variation diagram of the optimal solution after the regression equation is optimized by the neural network;
图5为最优组合的HFSS仿真结果图；Fig. 5 is the HFSS simulation result figure of optimum combination;
图6为最优组合的实验验证结果图。Figure 6 is a diagram of the experimental verification results of the optimal combination.
具体实施方式detailed description
下面结合附图和实施例对本发明做进一步阐述，但不是对本发明的限定。The present invention will be further described below in conjunction with the accompanying drawings and embodiments, but the present invention is not limited thereto.
实施例：Example:
基于响应曲面法和遗传算法优化PCB微带线结构的方法，具体包括如下步骤：The method for optimizing the PCB microstrip line structure based on the response surface method and the genetic algorithm specifically includes the following steps:
(1)建立HFSS的微带线仿真分析模型，模型基板尺寸如表1所示；(1) Establish the microstrip line simulation analysis model of HFSS, and the model substrate size is shown in Table 1;
(2)获得频率为1GHz—5GHz下的回波损耗S11以及插入损耗S21，如图1、2所示；(2) Obtain the return loss S11 and insertion loss S21 under the frequency of 1GHz-5GHz, as shown in Figures 1 and 2;
(3)获取影响微带线的影响因素为：基板厚度H1、微带线宽度t、微带线厚度H2以及基板介电常数E；分别对各个因素选取3个水平值，其因素水平表如表2所示；(3) Obtain the influencing factors affecting the microstrip line as follows: substrate thickness H1, microstrip line width t, microstrip line thickness H2, and substrate dielectric constant E; select three levels for each factor respectively, and the factor levels are as follows Shown in Table 2;
(4)采用BOX-Behnken的中心组合设计模型，有29组仿真模型水平组合，其中24组为分析因子，5组为零点因子，即参数水平组合相同，用于实验误差估计；(4) Using BOX-Behnken's central combination design model, there are 29 groups of simulation model level combinations, of which 24 groups are analysis factors, and 5 groups are zero-point factors, that is, the parameter level combinations are the same, and are used for experimental error estimation;
(5)根据微积分知识，任一函数都可由若干个多项式分段近似表示，因此在实际问题中，无论变量和结果间关系复杂程度如何，总可以用多项式回归来分析计算，由于本文设计变量为4个且变量与目标之间函数关系为非线性，结合表2的实验样本数，选用基于泰勒展开式的二阶多项式模型：(5) According to the knowledge of calculus, any function can be approximated by several polynomials. Therefore, in practical problems, no matter how complicated the relationship between variables and results is, polynomial regression can always be used to analyze and calculate. Because the design variables in this paper There are 4 variables and the functional relationship between the variables and the target is non-linear. Combined with the number of experimental samples in Table 2, a second-order polynomial model based on Taylor expansion is selected:
(A)式中包括常数项α0、线性项线性交叉项二次项αi为线性项系数；αij为线性交叉项系数；αii为二次项系数；ε为随机误差；x为设计变量；Y为目标值；n为变量个数。(A) formula includes constant term α0 , linear term linear cross term quadratic term αi is the coefficient of the linear term; αij is the coefficient of the linear cross term; αii is the coefficient of the quadratic term; ε is the random error; x is the design variable; Y is the target value; n is the number of variables.
(6)对表2中实验因子组合及其结果进行二次多元回归拟合，得到回波损耗(Y)对基板高度(X1)、微带线宽度(X2)、微带线厚度(X3)、介电常数(X4)的二次多项式回归方程为：(6) Perform quadratic multiple regression fitting on the experimental factor combinations and their results in Table 2 to obtain the return loss (Y) versus substrate height (X1 ), microstrip line width (X2 ), and microstrip line thickness ( The quadratic polynomial regression equation of X3 ), dielectric constant (X4 ) is:
(7)为了确保回归方程可信，对表2中数据进行了方差分析和模型的显著性验证，得到回归方程相关评价指标，结果如表3所示；(7) In order to ensure the credibility of the regression equation, variance analysis and model significance verification were carried out on the data in Table 2, and relevant evaluation indicators of the regression equation were obtained. The results are shown in Table 3;
(8)响应面分析得到的模型“Preb>F”小于0.0001，一般小于0.05即表示该项显著，即响应面模型回归效果特别明显；回归方程系数R-Squared为0.956，表明回归方程拟合度很高；回归方程调整系数Adj R-Squared为0.946，更准确的反映出方程的拟合度很高；回归方程预测系数Pred R-Squared为0.912，表明方程预测准确度良好；以上结果系数都表明(B)式能够高度拟合表2中的实验结果，故回归方程准确可信；(8) The model "Preb>F" obtained by the response surface analysis is less than 0.0001, and generally less than 0.05 means that the item is significant, that is, the regression effect of the response surface model is particularly obvious; the regression equation coefficient R-Squared is 0.956, indicating the fitting degree of the regression equation It is very high; the regression equation adjustment coefficient Adj R-Squared is 0.946, which more accurately reflects the high degree of fitting of the equation; the regression equation prediction coefficient Pred R-Squared is 0.912, indicating that the prediction accuracy of the equation is good; the above result coefficients all show that (B) formula can highly fit the experimental results in Table 2, so the regression equation is accurate and credible;
(9)利用遗传算法对上诉回归方程进行优化，该算法首先从定义域中随机确一组初始解，进而搜索领范围内目标函数的最优或算法首先从定义域中随机确一组初始解，进而搜索领范围内目标函数的最优或次优解；(9) Use the genetic algorithm to optimize the appeal regression equation. The algorithm first randomly determines a set of initial solutions from the domain of definition, and then searches for the optimum of the objective function in the domain or the algorithm first randomly determines a set of initial solutions from the domain of definition , and then search for the optimal or suboptimal solution of the objective function within the range;
所述的遗传算法优化回归方程，具体如下步骤：The genetic algorithm optimizes the regression equation, specifically as follows:
步骤a：采用随机方式生成初始种群；Step a: Randomly generate the initial population;
步骤b：获得当前进化代数gen和最优适应度值；Step b: Obtain the current evolution algebra gen and optimal fitness value;
步骤c：分别对种群实施交叉操作；Step c: Carry out the cross operation on the population respectively;
步骤d：分别对种群实施变异操作；Step d: Carry out mutation operations on the populations respectively;
步骤e：分别对种群实施进化逆转；Step e: implement evolutionary reversal on the populations respectively;
步骤f：将种群作为整体计算适应度函数值，并采用最优保存策略选择最佳个体；Step f: Calculate the fitness function value of the population as a whole, and use the optimal preservation strategy to select the best individual;
步骤g：种群更新后重新判断，若gen值小于50且num值大于0，则对种群实施局部灾变，然后返回步骤b，否则直接返回步骤b；算法的最大遗传代数设为50代，gen值超过50则终止进化。Step g: re-judgment after the population is updated, if the gen value is less than 50 and the num value is greater than 0, implement a local catastrophe on the population, and then return to step b, otherwise directly return to step b; the maximum genetic generation of the algorithm is set to 50 generations, and the gen value If it exceeds 50, the evolution will be terminated.
(10)通过MATLAB遗传算法工具箱以回拨损耗S11最低为目标进行参数优化；问题均值和最优解变化如图3、图4所示。(10) Through the MATLAB genetic algorithm toolbox, the parameters are optimized with the goal of the minimum callback loss S11; the average value of the problem and the change of the optimal solution are shown in Figure 3 and Figure 4.
(11)根据上诉因素参数表里设定影响因子的取值范围，获得最优组合为PCB基板H1为0.5mm,微带线宽度为0.4mm,微带线厚度为0.075mm，介电常数E为4.4，此时获得5GHZ预测回波损耗值为-13.006dB。(11) According to the value range of the influence factor set in the appeal factor parameter table, the optimal combination is obtained as the PCB substrate H1 is 0.5mm, the width of the microstrip line is 0.4mm, the thickness of the microstrip line is 0.075mm, and the dielectric constant E is 4.4, and the 5GHZ predicted return loss value is -13.006dB.
(12)根据上述所获得最后参数组合，建立相应的HFSS微带线仿真模型，其仿真结果曲线如图5所示，其5GHZ条件下的回波损耗值S11为-12.8dB，与遗传算法预测值极为接近，证明了遗传算法优化微带线结构的有效性。(12) According to the final parameter combination obtained above, a corresponding HFSS microstrip line simulation model is established, and the simulation result curve is shown in Figure 5, and the return loss value S11 under the 5GHZ condition is -12.8dB, which is consistent with the genetic algorithm prediction The values are very close, which proves the effectiveness of the genetic algorithm to optimize the microstrip line structure.
(13)根据上述所获得最优参数组合，制作最优参数组合的试验样件，测量得到实测曲线图大致趋势以及回波损耗值S11较为接近，如图6所示，证实了基于HFSS的微带线仿真模型的正确性已经基于响应曲面-遗传算法的准确性。(13) According to the optimal parameter combination obtained above, the test sample of the optimal parameter combination is made, and the general trend of the measured curve and the return loss value S11 are relatively close to each other, as shown in Figure 6, which confirms that the HFSS-based micro The correctness of the stripline simulation model has been based on the accuracy of the response surface-genetic algorithm.
表1模型尺寸图Table 1 Model Dimensions
表2因素水平表Table 2 Factor level table
表3 29组参数组合结果Table 3 Combination results of 29 groups of parameters
表4响应面分析结果Table 4 Response surface analysis results
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710022783.7ACN106844924B (en) | 2017-01-12 | 2017-01-12 | Method for optimizing PCB microstrip line structure based on response surface method and genetic algorithm |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710022783.7ACN106844924B (en) | 2017-01-12 | 2017-01-12 | Method for optimizing PCB microstrip line structure based on response surface method and genetic algorithm |
| Publication Number | Publication Date |
|---|---|
| CN106844924Atrue CN106844924A (en) | 2017-06-13 |
| CN106844924B CN106844924B (en) | 2021-02-23 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710022783.7AActiveCN106844924B (en) | 2017-01-12 | 2017-01-12 | Method for optimizing PCB microstrip line structure based on response surface method and genetic algorithm |
| Country | Link |
|---|---|
| CN (1) | CN106844924B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107832526A (en)* | 2017-11-09 | 2018-03-23 | 桂林电子科技大学 | A kind of method of optimization BGA solder joint return losses |
| CN108829937A (en)* | 2018-05-24 | 2018-11-16 | 郑州云海信息技术有限公司 | A method of optimization PCB high speed signal via hole |
| CN109002644A (en)* | 2018-08-10 | 2018-12-14 | 桂林电子科技大学 | A kind of optimization method of multi-chip module fluid channel radiator structure |
| CN109063298A (en)* | 2018-07-23 | 2018-12-21 | 桂林电子科技大学 | A kind of structure parameter optimizing method improving fluid channel heat dissipation performance |
| CN109190152A (en)* | 2018-07-23 | 2019-01-11 | 桂林电子科技大学 | A kind of CSP welding spot structure parameter optimization method reducing the stress under power cycle-Harmony response coupling |
| CN109376372A (en)* | 2018-08-29 | 2019-02-22 | 桂林电子科技大学 | A method for optimizing post-solder coupling efficiency at key locations of optical interconnect modules |
| CN111159921A (en)* | 2020-01-17 | 2020-05-15 | 安徽瑞迪微电子有限公司 | A design method of IGBT |
| CN112016243A (en)* | 2020-07-30 | 2020-12-01 | 东南大学 | Traffic flow prediction model parameter calibration method based on response surface |
| CN112257373A (en)* | 2020-11-13 | 2021-01-22 | 江苏科技大学 | Snake-shaped PCB antenna return loss prediction method based on three-body training algorithm |
| CN113051856A (en)* | 2021-03-24 | 2021-06-29 | 西安电子科技大学 | Optimization method of integrated circuit design data |
| CN113095677A (en)* | 2021-04-13 | 2021-07-09 | 北京工业大学 | Machining process quantitative control method based on reverse derivation of machining quality |
| CN114781186A (en)* | 2022-05-26 | 2022-07-22 | 江南大学 | Method for simulating and predicting energy consumption in cutting process of intelligent processing production line |
| US20230016096A1 (en)* | 2021-07-15 | 2023-01-19 | Montage Electronics (Shanghai) Co., Ltd. | Method for obtaining board parameters of printed circuit board |
| CN116298434A (en)* | 2023-02-24 | 2023-06-23 | 珠海市一博科技有限公司 | SOCKET probe impedance analysis method |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030126562A1 (en)* | 2001-12-27 | 2003-07-03 | Hamlin Christopher L. | System and method for coevolutionary circuit design |
| GB2479662A (en)* | 2006-01-17 | 2011-10-19 | Halliburton Energy Serv Inc | A control method predicting a flow rate between a reservoir and a wellbore |
| CN103745029A (en)* | 2013-12-12 | 2014-04-23 | 西安电子工程研究所 | High-power isolation resistor of microwave power divider and design method of isolation resistor |
| CN106096125A (en)* | 2016-06-08 | 2016-11-09 | 湖南大学 | One not uniform thickness tailor welded weld line optimization method |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030126562A1 (en)* | 2001-12-27 | 2003-07-03 | Hamlin Christopher L. | System and method for coevolutionary circuit design |
| GB2479662A (en)* | 2006-01-17 | 2011-10-19 | Halliburton Energy Serv Inc | A control method predicting a flow rate between a reservoir and a wellbore |
| CN103745029A (en)* | 2013-12-12 | 2014-04-23 | 西安电子工程研究所 | High-power isolation resistor of microwave power divider and design method of isolation resistor |
| CN106096125A (en)* | 2016-06-08 | 2016-11-09 | 湖南大学 | One not uniform thickness tailor welded weld line optimization method |
| Title |
|---|
| LEI LIU等: ""Numerical Techniques for Multi-Objective Synthesis of an Inverted-S Antenna"", 《PROCEEDINGS OF THE 17TH CONFERENCE ON THE COMPUTATION OF ELECTROMAGNETIC FIELDS》* |
| 傅厦龙等: ""基于响应曲面和遗传算法的工艺参数优化"", 《高分子材料科学与工程》* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107832526A (en)* | 2017-11-09 | 2018-03-23 | 桂林电子科技大学 | A kind of method of optimization BGA solder joint return losses |
| CN107832526B (en)* | 2017-11-09 | 2020-11-17 | 桂林电子科技大学 | Method for optimizing return loss of BGA welding spot |
| CN108829937B (en)* | 2018-05-24 | 2022-02-18 | 郑州云海信息技术有限公司 | Method for optimizing PCB high-speed signal via hole |
| CN108829937A (en)* | 2018-05-24 | 2018-11-16 | 郑州云海信息技术有限公司 | A method of optimization PCB high speed signal via hole |
| CN109063298A (en)* | 2018-07-23 | 2018-12-21 | 桂林电子科技大学 | A kind of structure parameter optimizing method improving fluid channel heat dissipation performance |
| CN109190152A (en)* | 2018-07-23 | 2019-01-11 | 桂林电子科技大学 | A kind of CSP welding spot structure parameter optimization method reducing the stress under power cycle-Harmony response coupling |
| CN109190152B (en)* | 2018-07-23 | 2023-04-07 | 桂林电子科技大学 | CSP welding spot structural parameter optimization method for reducing stress under power cycle-harmonic response coupling |
| CN109002644A (en)* | 2018-08-10 | 2018-12-14 | 桂林电子科技大学 | A kind of optimization method of multi-chip module fluid channel radiator structure |
| CN109376372B (en)* | 2018-08-29 | 2022-11-18 | 桂林电子科技大学 | Method for optimizing postweld coupling efficiency of key position of optical interconnection module |
| CN109376372A (en)* | 2018-08-29 | 2019-02-22 | 桂林电子科技大学 | A method for optimizing post-solder coupling efficiency at key locations of optical interconnect modules |
| CN111159921B (en)* | 2020-01-17 | 2023-06-16 | 安徽瑞迪微电子有限公司 | IGBT design method |
| CN111159921A (en)* | 2020-01-17 | 2020-05-15 | 安徽瑞迪微电子有限公司 | A design method of IGBT |
| CN112016243A (en)* | 2020-07-30 | 2020-12-01 | 东南大学 | Traffic flow prediction model parameter calibration method based on response surface |
| CN112257373B (en)* | 2020-11-13 | 2022-05-17 | 江苏科技大学 | Prediction method of return loss of serpentine PCB antenna based on three-body training algorithm |
| CN112257373A (en)* | 2020-11-13 | 2021-01-22 | 江苏科技大学 | Snake-shaped PCB antenna return loss prediction method based on three-body training algorithm |
| CN113051856A (en)* | 2021-03-24 | 2021-06-29 | 西安电子科技大学 | Optimization method of integrated circuit design data |
| CN113051856B (en)* | 2021-03-24 | 2024-01-19 | 西安电子科技大学 | Optimization method for integrated circuit design data |
| CN113095677A (en)* | 2021-04-13 | 2021-07-09 | 北京工业大学 | Machining process quantitative control method based on reverse derivation of machining quality |
| US20230016096A1 (en)* | 2021-07-15 | 2023-01-19 | Montage Electronics (Shanghai) Co., Ltd. | Method for obtaining board parameters of printed circuit board |
| US11782086B2 (en)* | 2021-07-15 | 2023-10-10 | Montage Electronics (Shanghai) Co., Ltd. | Method for obtaining board parameters of printed circuit board |
| CN114781186A (en)* | 2022-05-26 | 2022-07-22 | 江南大学 | Method for simulating and predicting energy consumption in cutting process of intelligent processing production line |
| CN114781186B (en)* | 2022-05-26 | 2025-05-27 | 江南大学 | A method for simulating and predicting energy consumption in cutting process of intelligent machining production line |
| CN116298434A (en)* | 2023-02-24 | 2023-06-23 | 珠海市一博科技有限公司 | SOCKET probe impedance analysis method |
| Publication number | Publication date |
|---|---|
| CN106844924B (en) | 2021-02-23 |
| Publication | Publication Date | Title |
|---|---|---|
| CN106844924B (en) | Method for optimizing PCB microstrip line structure based on response surface method and genetic algorithm | |
| CN107832526B (en) | Method for optimizing return loss of BGA welding spot | |
| CN114239389B (en) | A high-fidelity modeling method and system for a microwave radio frequency process IP simulation model | |
| CN114065607B (en) | Offshore non-uniform evaporation waveguide profile inversion method based on deep convolutional network | |
| NL2036175B1 (en) | Method and apparatus for coupling superconducting qubit, electronic device, computer medium | |
| CN110728034A (en) | A Fast Multi-Object Modeling Method for Antennas Using Multi-level Collaborative Machine Learning | |
| Müller et al. | Energy-aware signal integrity analysis for high-speed pcb links | |
| CN112257373B (en) | Prediction method of return loss of serpentine PCB antenna based on three-body training algorithm | |
| Che et al. | Investigation of segmentation method for enhancing high frequency simulation accuracy of Q3D extractor | |
| Kumar et al. | Chip-to-chip copper interconnects with rough surfaces: Analytical models for parameter extraction and performance evaluation | |
| Wang et al. | Multi‐harmonic sources location based on sparse component analysis and complex independent component analysis | |
| Sánchez-Masís et al. | FNNs models for regression of S-parameters in multilayer interconnects with different electrical lengths | |
| Sahu et al. | An automated machine learning approach to inkjet printed component analysis: A step toward smart additive manufacturing | |
| CN103474737B (en) | Millimeter wave E-plane filter and membrane modeling method based on support vector machine for membrane modeling | |
| Roy et al. | Preliminary application of deep learning to design space exploration | |
| CN115455883A (en) | Method for establishing scalable model of grounded coplanar waveguide based on BP neural network | |
| Zhou et al. | Simulation and measurement for shielding effectiveness of small size metal enclosure | |
| Koziel et al. | Modeling of microwave devices with space mapping and radial basis functions | |
| Liu et al. | A GA-BPNN collaborative method based on high-quality datasets for designing and optimizing via-holes vertical transition | |
| Wang et al. | Studying the effect of drilling uncertainty on signal propagation through vias | |
| Preibisch et al. | Sensitivity analysis of via impedance using polynomial chaos expansion | |
| CN109376372B (en) | Method for optimizing postweld coupling efficiency of key position of optical interconnection module | |
| Liu et al. | Data-driven electronic packaging structure inverse design with an adaptive surrogate model | |
| CN103136397B (en) | A kind of method obtaining electromagnetic response curvilinear characteristic parameter and device thereof | |
| Zhang et al. | Efficient design optimization of microwave circuits using parallel computational methods |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| EE01 | Entry into force of recordation of patent licensing contract | Application publication date:20170613 Assignee:Guilin Gaopu Electronic Technology Co.,Ltd. Assignor:GUILIN University OF ELECTRONIC TECHNOLOGY Contract record no.:X2022450000412 Denomination of invention:Optimization of PCB Microstrip Line Structure Based on Response Surface Methodology and Genetic Algorithm Granted publication date:20210223 License type:Common License Record date:20221227 | |
| EE01 | Entry into force of recordation of patent licensing contract | ||
| EE01 | Entry into force of recordation of patent licensing contract | Application publication date:20170613 Assignee:Guilin Yuanjing Electronic Technology Co.,Ltd. Assignor:GUILIN University OF ELECTRONIC TECHNOLOGY Contract record no.:X2024980028570 Denomination of invention:Method for optimizing PCB microstrip line structure based on response surface methodology and genetic algorithm Granted publication date:20210223 License type:Common License Record date:20241202 | |
| EE01 | Entry into force of recordation of patent licensing contract |