基于频率下降率的结构损伤自适应神经网络识别

笔者探讨了动量系数和学习率自适应调整的神经网络算法及结构裂纹损伤识别特征参数的选取,提出以反映结构损伤位置和程度的频率下降率作为结构裂纹损伤识别的特征参数,利用有限元网格细化法对结构裂纹损伤进行数值模拟,获取训练样本数据,通过自适应神经网络对结构裂纹损伤问题进行识别研究。从结构裂纹损伤识别实例的结果中可以看出,采用频率下降率和自适应神经网络技术对结构裂纹进行损伤识别分析具有较高的精度和可靠性。     

钢筋混凝土叠合梁斜截面承载力数值模拟与影响因素分析

钢筋混凝土叠合梁的斜截面承载力,从理论上定量分析存在着较多困难,但数值模拟分析是一种直接有效的途径。笔者对33根尺寸、配筋相同,叠合参数及混凝土强度不同的无腹筋叠合梁和1根整浇对比梁用ABAQUS进行了非线性有限元数值模拟,并且分析了叠合参数以及混凝土强度对叠合梁斜截面承载力的影响,给出了斜截面承载力的计算公式。该分析结果表明,叠合梁的斜截面承载力是整浇梁的1.00~1.14倍。所获分析结果对于促进叠合梁斜截面理论的深入研究、指导叠合梁在工程中的合理设计以及类似的数值模拟分析均有一定的参考价值。     

APPLICATION OF PARAMETER OPTIMIZATION METHOD FOR CALIBRATING TANK MODEL1

ABSTRACT: Many automatic calibration processes have been proposed to efficiently calibrate the 16 parameters involved in the four‐layered tank model. The Multistart Powell and Stuffed Complex Evolution (SCE) methods are considered the best two procedures. Two rainfall events were designed to compare the performance and efficiency of these two methods. The first rainfall event is short term and the second designed for long term rainfall data collection. Both rainfall events include a lengthy no‐rainfall period. Two sets of upper and lower values for the search range were selected for the numerical tests. The results show that the Multistart Powell and SCE methods are able to obtain the true values for the 16 parameters with a sufficiently long no‐rainfall period after a rainfall event. In addition, by using two selected objective functions, one based on root mean square error and one based on root mean square relative error criteria, it is found that the no‐rainfall period lengths necessary to obtain the converged true values for the 16 parameters are roughly the same. The SCE method provides a more efficient search based on an appropriate preliminary search range. The Multistart Powell method, on the other hand, leads to more accurate search results when there is no suitable search range selected based on the parameter calibration experience.     

PARAMETER IDENTIFICATION IN A TWO‐MULTIPLIER SEDIMENT YIELD MODEL1

ABSTRACT: A process based, distributed runoff erosion model (KINEROS2) was used to examine problems of parameter identification of sediment entrainment equations for small watersheds. Two multipliers were used to reflect the distributed nature of the sediment entrainment parameters: one multiplier for a raindrop induced entrainment parameter, and one multiplier for a flow induced entrainment parameter. The study was conducted in three parts. First, parameter identification was studied for simulated error free data sets where the parameter values were known. Second, the number of data points in the simulated sedigraphs was reduced to reflect the effect of temporal sampling frequency on parameter identification. Finally, event data from a small range‐land watershed were used to examine parameter identifiability when the parameter values are unknown. Results demonstrated that whereas unique multiplier values can be obtained for simulated error free data, unique parameter values could not be obtained for some event data. Unique multiplier values for raindrop induced entrainment and flow induced entrainment were found for events with greater than a two‐year return period (～25 mm) that also had at least 10 mm of rain in ten minutes. It was also found that the three‐minute sampling frequency used for the sediment sampler might be inadequate to identify parameters in some cases.     

EVALUATION OF METHODS FOR ESTIMATING STREAM WATER QUALITY PARAMETERS IN A TRANSIENT MODEL FROM STOCHASTIC DATA1

ABSTRACT: In this study the estimation of parameters in water quality models represented by linear first order partial differential equations is investigated. Two sets of simulated input-output data, one with input noise and the other with output measurement error, were used. The parameters were estimated by a gradient technique (Bard's method) and a pattern search technique. The results indicate that the output measurement error significantly affects the values of parameter estimates as compared to the noise added to the input. Bard's method consistently gave results with a smaller sum of square value.     

面源模式数值模拟计算分析

系统地介绍了面源模式计算方法.为了提高面源模式计算的准确性,提出了面源反扩散参数的概念.采用数值模拟计算的方法,对面源模式做了实际计算,并将其计算结果与虚点源后置法做了分析和比较.给出了采用虚点源后置法对面源内部和近源处计算结果的修正方法.并对虚点源后置法横风方向计算结果进行了修正.在计算过程中,令面源(y方向)内部所有计算点的浓度与中心点浓度相同,在面源两边0.45 L范围以外取实际浓度值.     

起重机主梁裂纹修补参数的研究

采用ANSYS有限元分析软件，对因出现裂纹的起重机主梁而添加的补强板进行系列计算，得出修补后主梁裂纹的应力及应力强度因子的变化规律，结果表明，只有采用适当参数的补强板，才可消除起重机主梁裂纹的扩展趋势，从而保证起重机的安全。     

株洲市霞湾污水处理厂设计

株洲市霞湾污水处理厂是为保护湘江而建的工程，设计规模为10万m^3/d，根据污水水质及排放要求，处理工艺采用了不设初沉池的氧化沟活性污泥法；污泥处理采用了不经消化直接脱水工艺。介绍各处理单元的主要设计参数和所选用设备的技术对参数。     

ASSESSMENT OF HYDRAULIC PROPERTIES IN AN UNCONFINED ALLUVIAL AQUIFER NEAR GRAND ISLAND,NEBRASKA1

ABSTRACT: A method is presented to analyze time-drawdown data from one or more observation wells for the calculation of four hydraulic parameters for unconfined aquifers: vertical hydraulic conductivity, horizontal hydraulic conductivity, storage coefficient, and specific yield. The hydraulic parameter results are further analyzed for reliability and the possible ranges of the actual parameter values. After verification using a theoretical example, the method was used to analyze pumping test data from 22 observation wells in an unconfined alluvial aquifer near Grand Island, Nebraska. Results indicate that this method can be used to efficiently calculate the four hydraulic parameters in this type of aquifer. The method can also identify the impact of measurement errors on the parameter estimates, and provide ranges of the actual parameter values. The parameter values calculated using this method were compared to those determined using other theories. It is found that this method is very useful for calculating the hydraulic properties from pumping test data and for analyzing the parameter reliability.     

阜山金矿成矿构造物理化学参量因子分析与成矿环境综合研究

采用因子分析方法 ,对影响阜山金矿成矿作用的各构造物理化学参量进行了相关分析 ,其目的是试图在传统定性研究的基础上 ,通过定量研究来探讨阜山金矿最佳成矿构造物理化学条件 ,进而建立一个初步的成矿构造物理化学模式。该模式指出 ,阜山金矿各构造物理化学参量的分布在矿区内是不均匀的 ,构造作用是控制成矿流体分布与状态的重要因素。在构造力松弛部位和构造力集中部位各参量分布场表现出不同的变化规律 ,两者之间存在一个构造物理化学界面 ,矿化主要发生在该界面附近构造力松弛部位。