人工蜂群优化支持向量机的电气故障诊断方法研究

摘要： 电气故障诊断具有重要的实际应用价值,针对电气故障诊断中的支持向量机（SVM）参数选择问题,提出了人工蜂群优化SVM的电气故障诊断模型。首先采用小波分析去除信号中的噪声,并提取特征,然后采用人工蜂群优化算法确定SVM的最优参数,建立电气故障诊断模型,最后通过与其他电气故障诊断模型进行对比实验。结果表明,WA-ABC-SVM可以描述电气设备状态与特征间的变化关系,提高了电气故障的诊断正确率,诊断结果要高于对比模型。     

基于GA优化支持向量机的变压器故障诊断

针对变压器故障诊断中出现的多故障分类问题,为提高支持向量机的多故障分类的准确率,利用遗传算法(GA)对支持向量机的相关参数进行了优化。将利用遗传算法优化的支持向量机(GA-SVM)应用于变压器故障诊断中,并与利用粒子群算法优化的支持向量机(PSO-SVM)的识别结果进行比较。对比试验结果可以看出,GA-SVM算法能够更为有效地选择支持向量机的相关参数,在很大程度上提高了变压器多故障分类的准确性。     

基于改进人工蜂群算法的锅炉NO_x排放预测优化

针对电厂循环流化床锅炉NOx排放问题进行了研究,并对人工蜂群算法进行了改进,结合最小二乘支持向量机建立了锅炉燃烧NOx排放模型,对锅炉可调参量进行了优化,降低了NOx排放浓度。将改进的人工蜂群算法与基本的人工蜂群算法和粒子群算法进行比较,说明基于改进人工蜂群算法所建立的模型能够很好的预测NOx的排放浓度,具有很强的辨识能力和泛化能力,同时也表明了改进人工蜂群算法计算速度快的优点及优化数据上的优势,通过仿真试验,优化后NOx排放浓度明显降低,体现了其工程实用价值。     

基于粗糙集与支持向量机的发动机故障诊断研究

在提取发动机气门机构故障特征的基础上,提出了采用粗糙集和支持向量机相结合的故障诊断方法。首先,基于粗糙集理论对故障诊断决策表进行属性约简,然后在最优决策属性的基础上使用支持向量机分类器对故障进行分类。实际诊断结果验证了采用粗糙集与支持向量机相结合的方法对故障进行诊断的可行性与有效性。     

基于主成分分析与支持向量机的汽轮机故障诊断

汽轮机故障诊断的一大难题是故障样本的缺乏,由于支持向量机针对小样本情况能取得很好的效果,为此,提出基于主成分分析与支持向量机的故障诊断方法,首先采用主成分分析方法对汽轮机故障数据进行故障特征提取,将特征向量作为支持向量分类器的输入,按照汽轮机的故障类型训练分类函数.对于支持向量机参数的选取,提出了基于错分样本数的蚁群优化算法.在小样本情况下对汽轮发电机组故障诊断进行了仿真研究.结果表明,应用该算法可以正确且有效地诊断多类汽轮机故障.     

基于粒子群优化支持向量机的变压器故障诊断

针对支持向量机(SVM)用于变压器故障诊断中模型参数具有不确定性的问题,采用粒子群优化(PSO)算法对支持向量机参数进行优化,减少了模型参数的不确定性。故障数据测试表明,PSO能快速、准确地优化SVM参数,二者的结合可有效完成变压器故障分类,并取得较为满意的效果。     

基于支持向量机的柴电混合动力故障诊断研究

对柴电混合动力系统级故障诊断进行了研究,利用仿真软件搭建了实时整车模型,并构建了基于支持向量机的柴电混合动力系统的诊断框架.采用一对一方法构建多分类器,故障识别准确率达到98％.构建了柴电混合动力系统故障诊断实时仿真平台,进行了基于支持向量机的柴电混合动力系统故障诊断实时仿真,验证了实时环境下基于支持向量机诊断算法能有...     

基于PSO优化LS-SVM的异步电动机振动故障诊断

为了提高异步电动机振动故障诊断的准确性,提出了基于粒子群算法优化最小二乘支持向量机的异步电动机振动故障诊断方法。先通过实验室对异步电动机各类故障的振动进行测试,对测试数据进行预处理,选择异步电动机不同位置振动信号的特征频率作为系统的输入,然后利用训练好的粒子群算法优化后的最小二乘支持向量机进行异步电动机振动的故障诊断。最终结果与其他诊断方法对比表明：该方法克服了样本训练时间较长并容易陷入局部收敛的缺点,同时诊断的准确率较高,有效地避免了异步电动机故障的误诊断。     

基于CLSSVM的风电机组齿轮箱故障诊断

针对风电机组齿轮箱传统故障诊断方法以全局误诊断率最小化为目标,忽略了误分类型之间的差别的问题,提出基于代价敏感最小二乘支持向量机(Cost-sensitive Least Squares Support Vector Machine,CLSSVM)的风电机组齿轮箱故障诊断方法。该方法在最小二乘支持向量机原始最优化问题中二次损失函数中嵌入不同样本的误分类代价,建立以误分类代价最小化为目标的CLSSVM故障诊断模型,并同最小二乘支持向量机和代价敏感支持向量机比较。实验结果表明,该方法能提高误分类代价高的故障类样本的诊断正确率,具有代价敏感性,其训练速度也足以满足风电机组齿轮箱故障诊断实时性的需求。     

基于全信息小波包和支持向量机的旋转机械故障诊断

针对传统旋转机械单通道故障诊断的信息不完整以及缺少故障样本等问题,提出了基于全信息小波包和支持向量机的旋转机械故障诊断方法.运用小波包频道能量分解技术提取了全信息能量特征向量,以此作为支持向量机多故障分类器的故障样本,经训练的分类器作为故障智能分类器可对设备工作状态进行自动识别和诊断.实验研究表明:基于全信息小波包和支持向量机的故障诊断方法能准确、有效地对旋转机械的工作状态和故障类型进行分类,显著提高了故障诊断的准确率.     

Natural-gas fueled spark-ignition (SI) and compression-ignition (CI) engine performance and emissions

Natural gas is a fossil fuel that has been used and investigated extensively for use in spark-ignition (SI) and compression-ignition (CI) engines. Compared with conventional gasoline engines, SI engines using natural gas can run at higher compression ratios, thus producing higher thermal efficiencies but also increased nitrogen oxide (NO_x) emissions, while producing lower emissions of carbon dioxide (CO₂), unburned hydrocarbons (HC) and carbon monoxide (CO). These engines also produce relatively less power than gasoline-fueled engines because of the convergence of one or more of three factors: a reduction in volumetric efficiency due to natural-gas injection in the intake manifold; the lower stoichiometric fuel/air ratio of natural gas compared to gasoline; and the lower equivalence ratio at which these engines may be run in order to reduce NO_x emissions. High NO_x emissions, especially at high loads, reduce with exhaust gas recirculation (EGR). However, EGR rates above a maximum value result in misfire and erratic engine operation. Hydrogen gas addition increases this EGR threshold significantly. In addition, hydrogen increases the flame speed of the natural gas-hydrogen mixture. Power levels can be increased with supercharging or turbocharging and intercooling. Natural gas is used to power CI engines via the dual-fuel mode, where a high-cetane fuel is injected along with the natural gas in order to provide a source of ignition for the charge. Thermal efficiency levels compared with normal diesel-fueled CI-engine operation are generally maintained with dual-fuel operation, and smoke levels are reduced significantly. At the same time, lower NO_x and CO₂ emissions, as well as higher HC and CO emissions compared with normal CI-engine operation at low and intermediate loads are recorded. These trends are caused by the low charge temperature and increased ignition delay, resulting in low combustion temperatures. Another factor is insufficient penetration and distribution of the pilot fuel in the charge, resulting in a lack of ignition centers. EGR admission at low and intermediate loads increases combustion temperatures, lowering unburned HC and CO emissions. Larger pilot fuel quantities at these load levels and hydrogen gas addition can also help increase combustion efficiency. Power output is lower at certain conditions than diesel-fueled engines, for reasons similar to those affecting power output of SI engines. In both cases the power output can be maintained with direct injection. Overall, natural gas can be used in both engine types; however further refinement and optimization of engines and fuel-injection systems is needed.     

Performance assessment of some ice TES systems

In this paper, a performance assessment of four main types of ice storage techniques for space cooling purposes, namely ice slurry systems, ice-on-coil systems (both internal and external melt), and encapsulated ice systems is conducted. A detailed analysis, coupled with a case study based on the literature data, follows. The ice making techniques are compared on the basis of energy and exergy performance criteria including charging, discharging and storage efficiencies, which make up the ice storage and retrieval process. Losses due to heat leakage and irreversibilities from entropy generation are included. A vapor-compression refrigeration cycle with R134a as the working fluid provides the cooling load, while the analysis is performed in both a full storage and partial storage process, with comparisons between these two. In the case of full storage, the energy efficiencies associated with the charging and discharging processes are well over 98% in all cases, while the exergy efficiencies ranged from 46% to 76% for the charging cycle and 18% to 24% for the discharging cycle. For the partial storage systems, all energy and exergy efficiencies were slightly less than that for full storage, due to the increasing effect wall heat leakage has on the decreased storage volume and load. The results show that energy analyses alone do not provide much useful insight into system behavior, since the vast majority of losses in all processes are a result of entropy generation which results from system irreversibilities.     

Decomposition of limestone in a solar reactor

The thermal decomposition of limestone has been selected as a model reaction for developing and testing an atmospheric open solar reactor. The reactor consists of a cyclone gas/particle separator which has been modified to let the concentrated solar energy enter through a windowless aperture. The reacting particles are directly exposed to the solar irradiation. Experimentation with a 60 kW reactor prototype was conducted at PSI's 90m² parabolic solar concentrator, in a continuous mode of operation. A counter-current flow heat exchanger was employed to preheat the reactants. Eighty five percent degree of calcination was obtained for cement raw material and 15% of the solar input was converted into chemical energy (enthalpy).The technical feasibility of the solar thermal decomposition of limestone was experimentally demonstrated. The use of solar energy as a source for high-temperature process heat offers the potential of reducing significantly the CO₂ emissions from lime producing plants. Such a solar thermochemical process can find application in sunny rural areas for avoiding deforestation.     

Effect of co-cultivation of Chlamydomonas reinhardtii with Azotobacter chroococcum on hydrogen production

Chlamydomonas reinhardtii cc124 and Azotobacter chroococcum bacteria were co-cultured with a series of volume ratios and under a variety of light densities to determine the optimal culture conditions and to investigate the mechanism by which co-cultivation improves H₂ yield. The results demonstrated that the optimal culture conditions for the highest H₂ production of the combined system were a 1:40 vol ratio of bacterial cultures to algal cultures under 200 μE m^?2 s^?1. Under these conditions, the maximal H₂ yield was 255 μmol mg^?1 Chl, which was approximately 15.9-fold of the control. The reasons for the improvement in H₂ yield included decreased O₂ content, enhanced algal growth, and increased H₂ase activity and starch content of the combined system.     

Biohydrogen production by Anabaena sp. PCC 7120 wild-type and mutants under different conditions: Light, nickel, propane, carbon dioxide and nitrogen

Increasing awareness of environmental problems caused by the current use of fossil fuel-based energy, has led to the search for alternatives. Hydrogen is a good alternative and the cyanobacterium Anabaena sp. PCC 7120 is naturally able to produce molecular hydrogen, photosynthetically from water and light. However, this H₂ is rapidly consumed by the uptake hydrogenase.This study evaluated the hydrogen production of Anabaena sp. PCC 7120 wild-type and mutants: hupL⁻ (deficient in the uptake hydrogenase), hoxH⁻ (deficient in the bidirectional hydrogenase) and hupL⁻/hoxH⁻ (deficient in both hydrogenases) on several experimental conditions, such as gas atmosphere (argon and propane with or without N₂ and/or CO₂ addition), light intensity (54 and 152 ??Em⁻²s⁻¹), light regime (continuous and light/dark cycles 16 h/8 h) and nickel concentrations in the culture medium.In every assay, the hupL⁻ and hupL⁻/hoxH⁻ mutants stood out over wild-type cells and the hoxH⁻ mutant. Nevertheless, the hupL⁻ mutant showed the best hydrogen production except in an argon atmosphere under 16 h light/8 h dark cycles at 54 ??Em⁻²s⁻¹ in the light period, with 1 ??M of NiCl₂ supplementation in the culture medium, and under a propane atmosphere.In all strains, higher light intensity leads to higher hydrogen production and if there is a daily 1% of CO₂ addition in the gas atmosphere, hydrogen production could increase 5.8 times, related to the great increase in heterocysts differentiation (5 times more, approximately), whereas nickel supplementation in the culture medium was not shown to increase hydrogen production. The daily incorporation of 1% of CO₂ plus 1% of N₂ did not affect positively hydrogen production rate.     

Investigation of the thermodynamic and kinetic properties of La–Fe–B system hydrogen-storage alloys

La–Fe–B hydrogen-storage alloys were prepared using a vacuum induction-quenching furnace with a rotating copper wheel. The thermodynamic and kinetic properties of the La–Fe–B hydrogen-storage alloys were investigated in this work. The P–C–I curves of the La–Fe–B alloys were measured over a H₂ pressure range of 10⁻³ MPa to 2.0 MPa at temperatures of 313, 328, 343 and 353 K. The P–C–I curves revealed that the maximum hydrogen-storage capacity of the alloys exceeded 1.23 wt% at a pressure of approximately 1.0 MPa and temperature of 313 K. The standard enthalpy of formation ΔH and standard entropy of formation ΔS for the alloys' hydrides, obtained according to the van't Hoff equation, were consistent with their application as anode materials in alkaline media. The alloys also exhibited good absorption/desorption kinetics at room temperature.     

Exergetic evaluation of 5 biowastes-to-biofuels routes via gasification

This paper presents the exergy analysis results for the production of several biofuels, i.e., SNG (synthetic natural gas), methanol, Fischer–Tropsch fuels, hydrogen, as well as heat and electricity, from several biowastes generated in the Dutch province of Friesland, selected as one of the typical European regions. Biowastes have been classified in 5 virtual streams according to their ultimate and proximate analysis. All production chains have been modeled in Aspen Plus in order to analyze their technical performance. The common steps for all the production chains are: pre-treatment, gasification, gas cleaning, water–gas-shift reactions, catalytic reactors, final gas separation and upgrading. Optionally a gas turbine and steam turbines are used to produce heat and electricity from unconverted gas and heat removal, respectively. The results show that, in terms of mass conversion, methanol production seems to be the most efficient process for all the biowastes. SNG synthesis is preferred when exergetic efficiency is the objective parameter, but hydrogen process is more efficient when the performance is analyzed by means of the 1st Law of Thermodynamics. The main exergy losses account for the gasification section, except in the electricity and heat production chain, where the combined cycle is less efficient.     

The Solar Office in context

The goal of sustainability in buildings can only hope to be realised if buildings are designed to both conserve and generate energy. The Solar Office at Doxford International is designed to minimise the use of energy while its external fabric is designed to replace such energy that is used. The recently completed building is now subject of a comprehensive monitoring programme. The programme covers both the performance of the 73 kW_p photovoltaic installation and the environmental conditions within the building as a whole. Hour by hour findings are posted on a dedicated web site. Photovoltaics could have the same impact on building form and layout as the invention of the passenger lift at the end of the last century.     

液压系统常见的故障诊断及处理

任何工程机械式液压设备使用时出现故障是不可避免的。但是怎样确定故障的原因及找到好的解决方法，这是使用者最关心的问题。讲述了液压系统常见的故障及其排除方法。     

Hydrogen recovery from the photovoltaic electroflocculation-flotation process for harvesting Chlorella pyrenoidosa microalgae

In this paper, an integrated process using photovoltaic power to harvest microalgae by electro-flocculation (EF) and hydrogen recovery is presented. It is mainly favorable in regions with high solar radiation. The electro-flocculation efficiency (EFE) of Chlorella pyrenoidosa microalgae was investigated using various types of electrodes (aluminum, iron, zinc, copper and a non-sacrificial electrode of carbon). The best results regarding the EFE, and biomass contamination were achieved with aluminum and carbon electrodes where the electrical energy demand of the process for harvesting 1 kg of algae biomass was 0.28 and 0.34 kWh, respectively, while the energy yield of harvested hydrogen was 0.052 and 0.005 kWh kg^?1, respectively. The highest harvesting efficiency of 95.83 ± 0.87% was obtained with the aluminum electrode.The experimental hydrogen yields obtained were comparable with those calculated from theory. With a low net energy demand, microalgae EF may be a useful and low-cost technology.