模糊预测控制在VAV空调系统中的应用

应用模糊控制的逻辑推理性能,借助于神经网络的学习能力,提出一种基于神经网络的模糊预测控制模型。并利用该模型对VAV(变风量)空调系统的提前开机时间进行预测控制,可获得较高的预测精度和较好的控制效果。     

基于模糊控制的单桩竖向极限承载力预测

为合理、准确地预测单桩竖向极限承载力,借助模糊神经网络较强的学习能力和模糊逻辑推理功能,建立了基于模拟退火算法的单桩竖向极限承载力预测模型,对收集到的样本进行了训练预测,并与常规的减法聚类算法预测结果进行了对比。分析结果表明:基于模拟退火算法的模糊神经网络模型预测单桩竖向极限承载力是可行有效的,具有较大的工程实用价值。     

基于SOFM神经网络模型的岩爆烈度等级预测方法

为简化岩爆烈度等级预测指标体系、解决预测分级模糊问题、利于工程人员分析预测结果,建立基于SOFM神经网络的岩爆烈度等级预测模型,并根据竞争层拓扑结构的不同将预测模型拓展成3个模型。将硐壁最大切向应力、岩石单轴抗压强度、岩石单轴抗拉强度作为模型输入向量,将40组国内外岩爆工程数据作为数据集输入3个模型进行训练、测试,3个模型在测试集上岩爆烈度等级预测正确率均达到90%。比较3个模型的聚类、测试及训练效果,得到竞争层神经元个数为16的预测模型最优。将最优预测模型的预测结果与可拓理论、Russenes判据、基于模糊C–均值算法粗糙集理论云模型的预测结果对比,基于SOFM神经网络模型的岩爆烈度等级预测方法优于其他方法,表明该方法具有一定可行性和实用性,为岩爆预测提供了一种新的方法和手段。     

改进的模糊Modular神经网络在既有建筑可靠性鉴定中的应用

在Takagi-Sugeno模糊逻辑系统的基础上,提出了改进的模糊Modular神经网络模型(IFMNN),并将该模型应用于既有建筑的可靠性鉴定.改进的模型是将传统的模糊Modular神经网络模型中的单输出改进为多输出.这种改进的多输入多输出的模糊Modular神经网络模型具有预测性能好、训练学习速度快的优点,它的系统门网络采用模糊C均值聚类算法代替K-means算法,专家网络的训练中引进了先进的Levenberg-Marquardt算法.在应用改进的模糊Modular神经网络模型对既有建筑进行可靠性鉴定的过程中,综合考虑了各主要因素对既有建筑可靠性鉴定等级的影响,并将经量化处理的影响因素作为网络的外部输入,将网络计算得到的4个输出值分别作为样本对应于不同可靠性等级的隶属度,建筑可靠性鉴定的最终评判等级为最大隶属度所对应的等级.训练和预测样本的计算结果证明了改进的模糊Modular神经网络模型在既有建筑可靠性鉴定中应用的可行性和有效性.     

一种基于补偿模糊神经网络的水质预测方法

在污水处理系统过程控制中,对水质变化规律进行预测是控制系统可靠、稳定运行的重要环节。介绍了基于模糊逻揖和神经网络的补偿神经网络(CFNN)及其学习算法,利用CFNN学习速度快、学习过程稳定、全局动态优化运算等特点,建立污水处理厂CFNN的水质预测模型。实例预测结果表明该模型对初始值的选择不敏感,具有很好的收敛性和预测精度,适合实际工程应用。     

基于神经网络预测误差修正的广义预测控制

针对建模误差对预测控制的影响,提出一种基于神经网络预测误差修正的广义预测控制算法.被控对象采用递推最小二乘法(RLS)辨识其线性模型;预测误差由前馈神经网络进行建模,并用Powell快速优化方法训练网络.该方法无需预先离线训练神经网络,即可投入闭环控制,并具有理想的稳定性、跟踪性能和鲁棒性.仿真结果表明该算法能够有效地控制复杂系统.     

基于模糊-RBF BP神经网络的天然气负荷预测

针对天然气负荷变化的特点,综合考虑各种影响因素,提出一种基于模糊-RBF P神经网络的天然气负荷预测模型,并编制了负荷预测软件。在RBF网络模型中采用了最新邻聚类算法,实现了网络结构和参数的双重调节,提高了训练的速度和预测的精度。将该软件应用于实例计算,并与相关模型结果比较,证明该模型可以更加快速准确地预测天然气的负荷值。     

基于粒子群神经网络的空调冷负荷短期预测

为提高空调冷负荷预测精度,本文提出了基于PSO-BP算法的神经网络模型。将PSO算法与BP神经网络相结合,对大型商场的空调样本数据进行冷负荷预测实验。结果表明,与BP神经预测算法相比,该算法的预测精度更高,运行速度更快。     

基于模糊神经网络的铣床热误差预测模型研究

当前,铣床主轴加工产品容易受到热误差的影响,造成产品精度下降。对此,采用模糊神经网络模型预测铣床主轴热误差,并对预测结果进行比较和分析。建立神经网络径向基函数的表达式,给出了模糊推理系统和控制规则,创建了模糊RBF神经网络预测模型,对铣床主轴进行热误差验证。结果显示:铣床主轴采用RBF神经网络模型预测误差较大,其Y轴和Z轴输出最大误差分别为5.9μm和7.1μm;铣床主轴采用模糊RBF神经网络模型预测误差较小,其Y轴和Z轴输出最大误差分别为3.5μm和2.9μm。同时,模糊RBF神经网络模型预测误差跳动幅度较小。采用模糊RBF神经网络预测模型,可以补偿铣床运行时产生的热误差,提高铣床主轴加工精度。     

深大基坑施工变形的智能预测与控制

基于模糊控制理论,结合智能控制思想,构建针对施工变形的预测控制系统,本文所论述的智能预测控制系统主要构成要件为神经网络预测器与模糊控制器,采用神经网络多步预测与时间窗口滚动技术,对深大基坑施工进行全过程、全方位、多角度滚动预测。该预测系统可实现对多步施工变形的一次性预测,实现了对基坑施工参数的及早调整与变形控制。并且基于此,结合模糊理论针对基坑施工变形开发模糊控制器,从而对施工过程进行了实时主动控制。     

A Wavelet‐Based Adaptive Pole Assignment Method for Structural Control

This article presents a time varying wavelet‐based pole assignment (WPA) method to control seismic vibrations in multi‐degree of freedom (MDOF) structural systems. The discrete wavelet transform is used to determine the energy content over the frequency band of the response in real time. The frequency content was implemented in the Big Bang–Big Crunch algorithm to update the optimum values of the closed‐loop poles of the structural system adaptively. To calculate optimum gain matrix, a robust pole placement algorithm was used. The gain matrix is calculated online based on response characteristic in real time and must not be calculated a priori (offline) choice. The WPA is tested on a 10‐story structural system subject to several historical ground motions. It is observed that the WPA has advantages in some design problems. Numerical examples illustrate that the proposed approach reduces the displacement response of the structure in real time more than conventional linear quadratic regulator (LQR) controller.     

试论地质资料及桩基检测对桩基施工与质量评价的影响

本文作者通过对钻孔灌桩施工的质量监督,发现地质资料及桩基检测是如何对桩基施工及质量评价产生影响的,并按有关技术要求提处理意见。     

环形混凝土电杆挠度测量不确定度分析

研究了不确定度在电杆力学试验中的应用,分别对标距测量、支点位移、荷载传感器误差进行了不确定度评定。通过设计建立数学模型,初步分析了各不确定度的来源,并计算了各影响因素的灵敏度系数,最后根据不确定度传播率合成,给出了不确定度分析结果。结果表明,在所有对电杆挠度测量的影响因素中,由于标距L、L2测量的灵敏度系数较小,其对挠度的影响可忽略。     

Olfactometric approach for the evaluation of citizens' exposure to industrial emissions in the city of Terni, Italy

Since several years odour nuisance is a serious environmental concern in the city of Terni, whose citizens are repeatedly lamenting the presence of malodours. This paper describes the olfactometric approach adopted for assessing the odour impact on the city of Terni, caused by the co-presence of three important industrial poles: the steel industry pole, the chemical pole and a third industrial pole, comprising different activities for the treatment of wastewaters and solid waste. The combination of analyses by dynamic olfactometry and dispersion modelling allowed the evaluation of the citizens' exposure to the industrial odours, resulting in both the quantification of emissions and the assessment of their impact on the territory. The overall odour emissions were estimated to be equal to 218,000 ou_E s^− 1: 51% from the steel industry pole, 29% from the chemical pole and 20% from the other plants, respectively. The simulation of the emission dispersion shows that the odour impact relevant to all three studied industrial poles is considerable, actually investing almost the whole city of Terni. The study results also enabled the identification of the most problematic odour sources, which turned out to be the primary emissions from the furnace for the steel industry pole, Treofan and the wastewater treatment plant for the chemical pole and the municipal wastewater treatment plant and the pulper incinerator as far as the other plants are concerned.     

Estimation of life expectancy of wood poles in electrical distribution networks

Wood poles have been commonly used to support electrical lines throughout Canada. The electrical networks in Ontario alone use over 2 million wood poles to support distribution lines. The aging and degradation of wood pole infrastructure built in 1950s and 1960s have potential to increase the risk of pole failure resulting in power outage and costly unplanned maintenance work. Therefore, the development of a scientific and cost-effective asset management philosophy has become increasingly important to power utilities.The paper presents a probabilistic approach to minimize the life-cycle cost of inspection and refurbishment of wood poles in a large distribution network. A key input of probabilistic methodology is the life-time distribution function of wood pole, though its estimation is often hampered by the lack of data. The paper presents a comprehensive statistical analysis and interpretation of actual wood pole inspection and surveillance data collected by power utilities. The statistical analysis provides estimates of life expectancy and the survival curve of a typical distribution wood pole in-service in the Canadian climate. The optimization of refurbishment policy presented in the paper is generic and equally applicable to the asset management of other civil infrastructure systems.     

Hydro-mechanical reinforcements of live poles to slope stability

Soil bioengineering using live poles is an environmentally friendly technique for shallow slope stabilisation. However, it remains unclear in this technique whether the hydrological effects of pole transpiration are significant to slope stabilisation, compared to mechanical reinforcement by structural poles and their fibrous roots. The aims of this study were to investigate the hydro-mechanical reinforcement effects of live poles and to evaluate their effectiveness for shallow slope stabilisation, giving due consideration to the different pole growth stages. Finite-element seepage-stability models were developed and validated against centrifuge model tests that investigated the rainfall-induced instability of a 45-degree clayey sand slope subjected to intense rainfall. The short-term stability right after the installation of the poles is critical because only structural poles, i.e., without fibrous root reinforcement or water uptake, are insufficient for reinforcement, even those as long as 2?m. Due to the absence of pole transpiration, positive pore water pressure of up to 10?kPa was built up near the slope toe, causing the significant mobilisation of shear strain and consequentially slope failure. In longer term, during which fibrous roots developed and provided additional mechanical reinforcement (via root cohesion) and transpiration-induced suction, no slope failure occurred due to the considerable amount of suction that was retained within the pole zone. It was mainly the pole transpiration before the rainfall, i.e., antecedent drying, that retained the suction, rather than the transpiration that took place during the rainfall.     

Finite element-based analyses of natural frequencies of long tapered hollow steel poles

Long tapered poles are commonly used to support Closed Circuit Television Cameras (CCTV) for security and traffic monitoring. Images received from CCTV are normally distorted due to the camera’s wind-induced vibration. To minimize the vibration experienced by the CCTV through development of a damping device to be placed at the interface between the pole and the camera, identification of the poles’ different frequencies as functions of poles’ geometric variables is essential. Thus, this study presents the development of natural frequency equations for long tapered hollow poles as functions of their geometric variables based on three-dimensional finite element model (FEM) analyses by taking into account the couplings between material, contact and geometric nonlinearities. The FEM results are compared and verified with the experimental data obtained from the instrumentation of a 60 ft pole. Sensitivity study is performed to identify the effect of different geometric parameters on the overall natural frequencies of the pole. Using the results from the parametric study, empirical formulae between the geometric parameters and the first, second and third mode natural frequencies for the long tapered hollow steel poles are obtained.     

Probabilistic design of prestressed concrete poles

A probability based procedure is presented for the optimum design of prestressed concrete poles. The cube strength of concrete, the ultimate strength of steel, the jacking stress at transfer, the cross sectional dimensions of the pole, the lever arm at which the load acts and the magnitude of the load acting on the pole are treated as random variables. The results obtained by the probabilistic design procedure are compared with those given by the deterministic procedure. The probability of failure of the pole obtained from the analytical method is found to be in good agreement with the value predicted by Monte Carlo simulation. The effect of variation of parameters like probability of failure and variability of cube strength of concrete is also studied. The optimum cost of the pole is found to increase as the probability of failure decreases. For large values of variability of concrete strength and small values of probability of failure, the compressive stress carrying capacity of the top section of the pole is found to be critical at the optimum point.     

Semi‐active algorithm for energy‐based predictive structural control using tuned mass dampers

A novel semi‐active control algorithm is developed and numerically evaluated for the suppression of undesirable structural vibrations. The mechanical energy of the vibrating structure is considered as the primary variable influencing the control action. This intuitive strategy is proposed to realize improved control of structural vibrations. The numerical study conducted reveals that the proposed energy‐based predictive (EBP) algorithm can be implemented on vibration control applications. The energy imparted to the structure is also reduced due to the proposed algorithm. The influence of the parameters of the proposed semi‐active tuned mass damper is studied. Further, the application of the proposed strategy on a realistic structure is numerically demonstrated by implementing the algorithm for the wind response control of a 76‐story benchmark building. The results show that the EBP algorithm is a competitive semi‐active strategy. The robustness of the strategy is also evaluated considering uncertainties in the properties of the benchmark building.     

碳纤维布加固钢筋混凝土电杆承载力试验研究

为了揭示采用碳纤维布(CFRP)加固钢筋混凝土电杆后的破坏机理和承载性能,完成4根混凝土电杆的试验研究,其中3根为碳纤维布加固的电杆和1根作对比的未加固电杆。通过试验观察各试件的受力全过程和破坏形态,获取荷载-位移全过程曲线和极限承载力等重要参数。研究结果表明:碳纤维布加固后,混凝土电杆的破坏过程没有明显预兆,破坏过程迅速,破坏形态表现为碳纤维被拉断;碳纤维布加固后,钢筋混凝土电杆的极限承载力有显著提高,提高的程度与碳纤维布粘贴层数有关,随着碳纤维布粘贴层数的增加,极限承载能力也跟着提高,但提高的程度与碳纤维布粘贴层数不呈比例。