Chapter

An Innovative Air Purification Method and Neural Network Algorithm Applied to Urban Streets

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

In the present work, multiphysics modeling was used to investigate the feasibility of a photocatalysis-based outdoor air purifying solution that could be used in high polluted streets, especially street canyons. The article focuses on the use of a semi-active photocatalysis in the surfaces of the street as a solution to remove anthropogenic pollutants from the air. The solution is based on lamellae arranged horizontally on the wall of the street, coated with a photocatalyst (TiO2), lightened with UV light, with a dimension of 8 cm × 48 cm × 1 m. Fans were used in the system to create airflow. A high purification percentage was obtained. An artificial neural network (ANN) was used to predict the optimal purification method based on previous simulations, to design purification strategies considering the energy cost. The ANN was used to forecast the amount of purified with a feed-forward neural network and a backpropagation algorithm to train the model.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

ResearchGate has not been able to resolve any citations for this publication.
Chapter
Full-text available
The advent of neural networks has led to the development of image classification algorithms that are applied to different fields. In order to recover the vital spatial factor parameters, for example, land cover and land utilization, image grouping is most important in remote sensing. Recently, benchmark classification accuracy was achieved using convolutional neural networks (CNNs) for land cover classification. The most well-known tool which indicates the presence of green vegetation from multispectral pictures is the Normalized Difference Vegetation Index (NDVI). This chaper utilizes the success of the NDVI for effective classification of a new satellite dataset, SAT-4, where the classes involved are types of vegetation. As NDVI calculations require only two bands of information, it takes advantage of both RED- and NIR-band information to classify different land cover. The number and size of filters affect the number of parameters in convolutional networks. Restricting the aggregate number of trainable parameters reduces the complexity of the function and accordingly decreases overfitting. The ConvNet Architecture with two band information, along with a reduced number of filters, was trained, and high-level features obtained from a tested model managed to classify different land cover classes in the dataset. The proposed architecture, results in the total reduction of trainable parameters, while retaining high accuracy, when compared with existing architecture, which uses four bands.
Article
Full-text available
Large Eddy Simulation (LES) undeniably has the potential to provide more accurate and more reliable results than simulations based on the Reynolds-averaged Navier-Stokes (RANS) approach. However, LES entails a higher simulation complexity and a much higher computational cost. In spite of some claims made in the past decades that LES would render RANS obsolete, RANS remains widely used in both research and engineering practice. This paper attempts to answer the questions why this is the case and whether this is justified, from the viewpoint of building simulation, both for outdoor and indoor applications. First, the governing equations and a brief overview of the history of LES and RANS are presented. Next, relevant highlights from some previous position papers on LES versus RANS are provided. Given their importance, the availability or unavailability of best practice guidelines is outlined. Subsequently, why RANS is still frequently used and whether this is justified or not is illustrated by examples for five application areas in building simulation: pedestrian-level wind comfort, near-field pollutant dispersion, urban thermal environment, natural ventilation of buildings and indoor airflow. It is shown that the answers vary depending on the application area but also depending on other—less obvious—parameters such as the building configuration under study. Finally, a discussion and conclusions including perspectives on the future of LES and RANS in building simulation are provided.
Article
Full-text available
Machine learning has proven to be a powerful technique during the past decades. Artificial neural network (ANN), as one of the most popular machine learning algorithms, has been widely applied to various areas. However, their applications for catalysis were not well-studied until recent decades. In this review, we aim to summarize the applications of ANNs for catalysis research reported in the literature. We show how this powerful technique helps people address the highly complicated problems and accelerate the progress of the catalysis community. From the perspectives of both experiment and theory, this review shows how ANNs can be effectively applied for catalysis prediction, the design of new catalysts, and the understanding of catalytic structures.
Article
Full-text available
Urban microclimate studies are gaining popularity due to rapid urbanization. Many studies documented that urban microclimate can affect building energy performance, human morbidity and mortality and thermal comfort. Historically, urban microclimate studies were conducted with observational methods such as field measurements. In the last decades, with the advances in computational resources, numerical simulation approaches have become increasingly popular. Nowadays, especially simulations with Computational Fluid Dynamics (CFD) is frequently used to assess urban microclimate. CFD can resolve the transfer of heat and mass and their interaction with individual obstacles such as buildings. Considering the rapid increase in CFD studies of urban microclimate, this paper provides a review of research reported in journal publications on this topic till the end of 2015. The studies are categorized based on the following characteristics: morphology of the urban area (generic versus real) and methodology (with or without validation study). In addition, the studies are categorized by specifying the considered urban settings/locations, simulation equations and models, target parameters and keywords. This review documents the increasing popularity of the research area over the years. Based on the data obtained concerning the urban location, target parameters and keywords, the historical development of the studies is discussed and future perspectives are provided. According to the results, early CFD microclimate studies were conducted for model development and later studies considered CFD approach as a predictive methodology. Later, with the established simulation setups, research efforts shifted to case studies. Recently, an increasing amount of studies focus on urban scale adaptation measures. The review hints a possible change in this trend as the results from CFD simulations can be linked up with different aspects (e.g. economy) and with different scales (e.g. buildings), and thus, CFD can play an important role in transferring urban climate knowledge into engineering and design practice.
Article
Full-text available
Photocatalytic concrete constitutes a promising technique to reduce a number of air contaminants such as NOx and VOC’s, especially at sites with a high level of pollution: highly trafficked canyon streets, road tunnels, the urban environment, etc. Ideally, the photocatalyst, titanium dioxide, is introduced in the top layer of the concrete pavement for best results. In addition, the combination of TiO2 with cement-based products offers some synergistic advantages, as the reaction products can be adsorbed at the surface and subsequently be washed away by rain. A first application has been studied by the Belgian Road Research Center (BRRC) on the side roads of a main entrance axis in Antwerp with the installation of 10.000 m² of photocatalytic concrete paving blocks. For now however, the translation of laboratory testing towards results in situ remains critical of demonstrating the effectiveness in large scale applications. Moreover, the durability of the air cleaning characteristic with time remains challenging for application in concrete roads. From this perspective, several new trial applications have been initiated in Belgium in recent years to assess the “real life” behavior, including a field site set up in the Leopold II tunnel of Brussels and the construction of new photocatalytic pavements on industrial zones in the cities of Wijnegem and Lier (province of Antwerp). This paper first gives a short overview of the photocatalytic principle applied in concrete, to continue with some main results of the laboratory research recognizing the important parameters that come into play. In addition, some of the methods and results, obtained for the existing application in Antwerp (2005) and during the implementation of the new realizations in Wijnegem and Lier (2010–2012) and in Brussels (2011–2013), will be presented.
Article
Full-text available
A body of evidence suggests that major changes involving the atmosphere and the climate, including global warming induced by anthropogenic factors, have impact on the biosphere and human environment. Studies on the effects of climate change on respiratory allergy are still lacking and current knowledge is provided by epidemiological and experimental studies on the relationship between allergic respiratory diseases, asthma and environmental factors, such as meteorological variables, airborne allergens, and air pollution. Urbanization with its high levels of vehicle emissions, and a westernized lifestyle are linked to the rising frequency of respiratory allergic diseases and bronchial asthma observed over recent decades in most industrialized countries. However, it is not easy to evaluate the impact of climate changes and air pollution on the prevalence of asthma in the general population and on the timing of asthma exacerbations, although the global rise in asthma prevalence and severity could also be an effect of air pollution and climate change. Since airborne allergens and air pollutants are frequently increased contemporaneously in the atmosphere, an enhanced IgE-mediated response to aeroallergens and enhanced airway inflammation could account for the increasing frequency of respiratory allergy and asthma in atopic subjects in the last 5 decades. Pollen allergy is frequently used to study the relationship between air pollution and respiratory allergic diseases, such as rhinitis and bronchial asthma. Epidemiologic studies have demonstrated that urbanization, high levels of vehicle emissions, and westernized lifestyle are correlated with an increased frequency of respiratory allergy prevalently in people who live in urban areas in comparison with people living in rural areas. Climatic factors (temperature, wind speed, humidity, thunderstorms, etc.) can affect both components (biological and chemical) of this interaction.
Article
Full-text available
Information on pedestrian-level wind (PLW) speed for wind comfort assessment can be obtained by wind-tunnel measurements or Computational Fluid Dynamics (CFD) simulations. Wind-tunnel measurements for PLW are routinely performed with low-cost techniques such as hot-wire or hot-film anemometers, Irwin probes or sand erosion, while Laser-Doppler Anemometry (LDA) and Particle-Image Velocimetry (PIV) are less often used because they are more expensive. CFD simulations are routinely performed by the relatively low-cost steady Reynolds-Averaged Navier-Stokes (RANS) approach. Large-Eddy Simulation (LES) is less often used because of its larger complexity and cost. This paper reviews wind-tunnel and CFD techniques to determine PLW speeds expressed generally in terms of amplification factors defined as the ratio of local mean wind speed to mean wind speed at the same position without buildings present. Some comparative studies systematically indicate that the low-cost wind-tunnel techniques and steady RANS simulations can provide accurate results (∼10%) at high amplification factors (> 1) while their accuracy can deteriorate at lower amplification factors (< 1). This does not necessarily compromise the accuracy of PLW comfort assessment, because the higher amplification factors provide the largest contribution to the discomfort exceedance probability in the comfort criterion. Although LDA, PIV and LES are inherently more accurate techniques, this paper supports the continued use of faster and more inexpensive techniques for PLW studies. Extrapolating a previous saying, we argue that pedestrian-level wind comfort is one of the few topics in wind engineering where nature is kind to us concerning turbulent flows.
Article
Full-text available
Despite past improvements in air quality, very large parts of the population in urban areas breathe air that does not meet European standards let alone the health-based World Health Organisation Air Quality Guidelines. Over the last 10 years, there has been a substantial increase in findings that particulate matter (PM) air pollution is not only exerting a greater impact on established health endpoints, but is also associated with a broader number of disease outcomes. Data strongly suggest that effects have no threshold within the studied range of ambient concentrations, can occur at levels close to PM2.5 background concentrations and that they follow a mostly linear concentration-response function. Having firmly established this significant public health problem, there has been an enormous effort to identify what it is in ambient PM that affects health and to understand the underlying biological basis of toxicity by identifying mechanistic pathways-information that in turn will inform policy makers how best to legislate for cleaner air. Another intervention in moving towards a healthier environment depends upon the achieving the right public attitude and behaviour by the use of optimal air pollution monitoring, forecasting and reporting that exploits increasingly sophisticated information systems. Improving air quality is a considerable but not an intractable challenge. Translating the correct scientific evidence into bold, realistic and effective policies undisputedly has the potential to reduce air pollution so that it no longer poses a damaging and costly toll on public health.
Article
Full-text available
This paper examines the forecasting performance of ARIMA and artificial neural networks model with published stock data obtained from New York Stock Exchange. The empirical results obtained reveal the superiority of neural networks model over ARIMA model. The findings further resolve and clarify contradictory opinions reported in literature over the superiority of neural networks and ARIMA model and vice versa.
Article
Full-text available
Computational Fluid Dynamics (CFD) model simulations of urban boundary layers have improved so that they are useful in many types of flow and dispersion analyses. The study described here is intended to assist in planning emergency response activities related to releases of chemical or biological agents into the atmosphere in large cities such as New York City. Five CFD models (CFD-Urban, FLACS, FEM3MP, FEFLO-Urban, and Fluent-Urban) have been applied by five independent groups to the same 3-D building data and geographic domain in Manhattan, using approximately the same wind input conditions. Wind flow observations are available from the Madison Square Garden March 2005 (MSG05) field experiment. It is seen from the many side-by-side comparison plots that the CFD models simulations of near-surface wind fields generally agree with each other and with field observations, within typical atmospheric uncertainties of a factor of two. The qualitative results shown here suggest, for example, that transport of a release at street level in a large city could reach a few blocks in the upwind and crosswind directions. There are still key differences seen among the models for certain parts of the domain. Further quantitative examinations of differences among the models and the observations are necessary to understand causal relationships.
Article
Full-text available
A review of patents on the application of titanium dioxide photocatalysis for air treatment is presented. A comparison between water treatment and air treatment reveals that the number of scientific publications dedicated to photocatalytic air treatment is significantly lower than the number of scientific manuscripts dedicated to photocatalytic water treatment, yet the situation is reversed upon comparing relevant patents. This indicates a growing interest in the implementation of photocatalysis for air treatment purposes, which surpasses that of water treatment.This manuscript analyzes the various patents in the area of air treatment, while differentiating between indoor air treatment and outdoor air treatment. Specific efforts were made to characterize the main challenges and achievements en-route for successful implementation, which were categorized according to mass transport, adsorption of contaminants, quantum efficiency, deactivation, and, no less important, the adherence and the long term stability of the photocatalyst.
Article
Full-text available
China’s coastal cities are experiencing rapid urbanization, which has resulted in many challenges. This paper presents a comprehensive index system for assessment of the level of urbanization based on four aspects: demographic urbanization, economic urbanization, social urbanization and spatial urbanization. The developed index system also characterizes the environment based on three factors: environmental pressure, environmental level and environmental control. Furthermore, a coupling coordination degree model (CCDM) focusing on the degree of coordination between urbanization and the environment was established using panel data collected from 2000 to 2008 for Lianyungang, China. The results showed that: (1) the dynamic of coordination between urbanization and the environment showed a U-shaped curve, and both sub-systems evolved into a superior balance during rapid urbanization; (2) social urbanization and environmental control make the greatest contribution to the coupling system, indicating that they are the critical factors to consider when adjusting coordination development during decision-making; and (3) the two parameters (a-urbanization, b-environment) that have been widely used in previous studies had less of an effect on the coupling coordinated system than the other factors considered herein.
Article
Full-text available
Artificial neural networks (ANNs) are relatively new computational tools that have found extensive utilization in solving many complex real-world problems. The attractiveness of ANNs comes from their remarkable information processing characteristics pertinent mainly to nonlinearity, high parallelism, fault and noise tolerance, and learning and generalization capabilities. This paper aims to familiarize the reader with ANN-based computing (neurocomputing) and to serve as a useful companion practical guide and toolkit for the ANNs modeler along the course of ANN project development. The history of the evolution of neurocomputing and its relation to the field of neurobiology is briefly discussed. ANNs are compared to both expert systems and statistical regression and their advantages and limitations are outlined. A bird's eye review of the various types of ANNs and the related learning rules is presented, with special emphasis on backpropagation (BP) ANNs theory and design. A generalized methodology for developing successful ANNs projects from conceptualization, to design, to implementation, is described. The most common problems that BPANNs developers face during training are summarized in conjunction with possible causes and remedies. Finally, as a practical application, BPANNs were used to model the microbial growth curves of S. flexneri. The developed model was reasonably accurate in simulating both training and test time-dependent growth curves as affected by temperature and pH.
Article
Computational fluid dynamics predictions of urban flow are subject to several sources of uncertainty, such as the definition of the inflow boundary conditions or the turbulence model. Compared to Reynolds-averaged Navier-Stokes (RANS) simulations, large eddy simulations (LES) can reduce turbulence model uncertainty by resolving the turbulence down to scales in the inertial subrange, but the presence of other uncertainties will not be reduced. The objective of this study is to present an initial investigation of the relative importance of these different types of uncertainties by comparing urban flow predictions obtained using RANS and LES to field measurements. The simulations are designed to reproduce measurements performed during the Joint Urban 2003 field experiments. The time-averaged velocity measured at an upstream wind sensor is used to define the inflow boundary condition, and the results are compared to time-averaged measurements at 34 locations in the downtown area. For the turbulence kinetic energy, the LES is found to be more accurate than the RANS in 80% of the available high-frequency measurement locations. For the mean velocity field, this number reduces to 50% of all stations. Comparison of the LES results with a previous inflow uncertainty quantification study for RANS shows that locations where the LES is less accurate than the RANS correspond to locations where the RANS solution is highly sensitive to the inflow boundary conditions. This suggests that inflow uncertainties can be a dominant factor, and that their effect on LES results should be quantified to guarantee predictive capabilities.
Article
Inevitable presence of volatile organic compounds (VOCs) in indoor environment and their adverse impact on human health and productivity have encouraged the development of various technologies for air pollution remediation. Among these technologies, photocatalytic oxidation (PCO) is regarded as one of the most promising methods and has been the focus of many research works in the last two decades. Titanium dioxide (TiO2) is by far the most investigated photocatalyst for photocatalytic degradation of gaseous VOCs. This review article is intended to provide a comprehensive overview of the application of commercial TiO2 photocatalysts for removal of VOCs in air. First, the fundamentals of photocatalytic oxidation are briefly discussed and common TiO2-based photocatalysts are introduced. Then, the relations between the characteristics of photocatalysts (e.g. crystallinity, surface area and surface chemistry) and photocatalytic activity as well as the influence of key operating parameters on PCO processes are investigated. Afterwards, the reaction mechanisms and identified reaction intermediates/by-products for the most prevalent VOC families are reviewed. Finally, the paper discusses the deactivation of photocatalysts during PCO processes and some of the common regeneration techniques.
Article
Photocatalyst is needed for cleaner environment and a better quality of life. This fact leads an idea of more eco-compatible use of light. This approach could become an integral component of strategies to reduce indoor air pollutants through the use of photocatalysts as construction materials. In this study, we discuss the quality of indoor air, sources of pollutants in the indoor environment, and the photocatalysis process. Moreover, the different parameters and uses of different photocatalysts in concrete for indoor air purification, starting from the early research until the current research, are described, and the reaction mechanism of photocatalytic oxidation and sustainable construction by using photocatalysts are also reviewed.
Article
Air pollutants emitted from vehicles in street canyons may be reactive, undergoing mixing and chemical processing before escaping into the overlying atmosphere. The deterioration of air quality in street canyons occurs due to combined effects of proximate emission sources, dynamical processes (reduced dispersion) and chemical processes (evolution of reactive primary and formation of secondary pollutants). The coupling between dynamics and chemistry plays a major role in determining street canyon air quality, and numerical model approaches to represent this coupling are reviewed in this article. Dynamical processes can be represented by Computational Fluid Dynamics (CFD) techniques. The choice of CFD approach (mainly the Reynolds-Averaged Navier-Stokes (RANS) and Large-Eddy Simulation (LES) models) depends on the computational cost, the accuracy required and hence the application. Simplified parameterisations of the overall integrated effect of dynamics in street canyons provide capability to handle relatively complex chemistry in practical applications. Chemical processes are represented by a chemical mechanism, which describes mathematically the chemical removal and formation of primary and secondary species. Coupling between these aspects needs to accommodate transport, dispersion and chemical reactions for reactive pollutants, especially fast chemical reactions with time scales comparable to or shorter than those of typical turbulent eddies inside the street canyon. Different approaches to dynamical and chemical coupling have varying strengths, costs and levels of accuracy, which must be considered in their use for provision of reference information concerning urban canopy air pollution to stakeholders considering traffic and urban planning policies.
Article
Purpose of the review: Air pollution continues to be a major public health concern affecting nine out of 10 individuals living in urban areas worldwide. Exposure to air pollution is the ninth leading risk factor for cardiopulmonary mortality. The aim of this review is to examine the current literature for the most recent updates on health effects of specific air pollutants and their impact on asthma, chronic obstructive pulmonary disease, lung cancer, and respiratory infection. Recent findings: A total of 53 publications were reviewed to establish new insights as to how air pollution is associated with pulmonary morbidity and mortality. Considerable past evidence suggests that air pollution is an important factor that enhances pulmonary disease, while also causing greater harm in susceptible populations, such as children, the elderly, and those of low socio-economic status worldwide. Asthma, chronic obstructive pulmonary disease, lung cancer, and respiratory infections all seem to be exacerbated because of exposure to a variety of environmental air pollutants with the greatest effects because of particulate matter, ozone, and nitrogen oxides. New publications reviewed reaffirm these findings. Summary: Continued vigilance will be essential to lessen the effects of air pollution on human health and pulmonary disease. Cooperation at a multinational level will be required on the part of governments, industry, energy-based enterprises, and the public working together to solve our air quality issues at the local, national, and global level.
Article
Due to the challenge of climate and energy crisis, renewable energy generation including solar generation has experienced significant growth. Increasingly high penetration level of photovoltaic (PV) generation arises in smart grid. Solar power is intermittent and variable, as the solar source at the ground level is highly dependent on cloud cover variability, atmospheric aerosol levels, and other atmosphere parameters. The inherent variability of large-scale solar generation introduces significant challenges to smart grid energy management. Accurate forecasting of solar power/irradiance is critical to secure economic operation of the smart grid. This paper provides a comprehensive review of the theoretical forecasting methodologies for both solar resource and PV power. Applications of solar forecasting in energy management of smart grid are also investigated in detail.
Chapter
Inspired by the sophisticated functionality of human brains where hundreds of billions of interconnected neurons process information in parallel, researchers have successfully tried demonstrating certain levels of intelligence on silicon. Examples include language translation and pattern recognition software. While simulation of human consciousness and emotion is still in the realm of science fiction, we, in this chapter, consider artificial neural networks as universal function approximators. Especially, we introduce neural networks which are suited for time series forecasts.
Article
Deep neural nets with a large number of parameters are very powerful machine learning systems. However, overfitting is a serious problem in such networks. Large networks are also slow to use, making it difficult to deal with overfitting by combining the predictions of many different large neural nets at test time. Dropout is a technique for addressing this problem. The key idea is to randomly drop units (along with their connections) from the neural network during training. This prevents units from co-adapting too much. During training, dropout samples from an exponential number of different "thinned" networks. At test time, it is easy to approximate the effect of averaging the predictions of all these thinned networks by simply using a single unthinned network that has smaller weights. This significantly reduces overfitting and gives major improvements over other regularization methods. We show that dropout improves the performance of neural networks on supervised learning tasks in vision, speech recognition, document classification and computational biology, obtaining state-of-the-art results on many benchmark data sets. © 2014 Nitish Srivastava, Geoffrey Hinton, Alex Krizhevsky, Ilya Sutskever and Ruslan Salakhutdinov.
Article
A commercially available TiO2 type (Kronos vlp 7000), doped with carbon and having an extended spectrum of light absorption wavelength, is investigated for photocatalytic degradation of a representative air pollutant, namely acetaldehyde, under visible light. The modeling of the studied system was carried out including the proposal of kinetic expressions for the contaminant degradation and generation of main intermediates based on the reaction mechanism, the mass balances in the reactor, and the lamps' superficial emission model to evaluate the radiation distribution in the reaction space. The predicted and experimental outlet concentrations of acetaldehyde and formaldehyde as main intermediate were found to be in good agreement obtaining a root mean square error equal to 13 %.
Article
This work evaluated the catalytic activity of TiO2 synthesized by the Pechini method. with varying molar ratios of 2:1, 3:1 and 4:1 of citric acid/metallic cations, in the photocatalytic degradation of methyl red dye in aqueous solution. The samples were characterized by X-ray diffraction, phase quantification by Rietveld structure refinement, and textural analysis by nitrogen adsorption, and their photocatalytic performance was bench- tested. The results indicated that the 3:1 and 4:1 samples contained two phases, with 84.4 and 89% of anatase phase and 15.6 and 11% of rutile phase, respectively. The 2:1 sample contained only anatase phase. The total discoloration of methyl red dye in 24 hours confirmed the high photocatalytic efficiency of the 2:1 sample, which was ascribed to the formation of monophasic anatase.
Article
The photocatalytic degradation of NOx in the gas phase was investigated comparing several commercial TiO2 sold as pigmentary-powders and characterized by crystallite sizes ranging from nano to micrometer dimensions. In particular the photocatalytic activity of the micro-sized sample was evaluated in comparison with the well-known activity of the nano-sized samples, being these last photocatalysts potentially dangerous due to the risk towards the human safety. The studied samples were precisely chosen among different commercially available products on the basis of the following features: pure anatase, uncoated surface, undoped material, not sold as photocatalytic materials. All samples reveal good photoactivity in the photodegradation of NOx in gas phase with an evident superiority of the nano-sized sample. However, the gap of activity between nano and micro-sized samples tends to be canceled when the starting NO2 concentration was reduced and fixed from 1000 to 200 ppb, a precise amount that is the first alert threshold for NO2 in air (World Health Organization). A proper kinetic model, based on the Langmuir–Hinshelwood mechanism and on the hypothesis of irreversible adsorption of the products on the catalysts surface, has been developed and discussed.
Article
This paper investigates how to train a recurrent neural network (RNN) using the Levenberg-Marquardt (LM) algorithm as well as how to implement optimal control of a grid-connected converter (GCC) using an RNN. To successfully and efficiently train an RNN using the LM algorithm, a new forward accumulation through time (FATT) algorithm is proposed to calculate the Jacobian matrix required by the LM algorithm. This paper explores how to incorporate FATT into the LM algorithm. The results show that the combination of the LM and FATT algorithms trains RNNs better than the conventional backpropagation through time algorithm. This paper presents an analytical study on the optimal control of GCCs, including theoretically ideal optimal and suboptimal controllers. To overcome the inapplicability of the optimal GCC controller under practical conditions, a new RNN controller with an improved input structure is proposed to approximate the ideal optimal controller. The performance of an ideal optimal controller and a well-trained RNN controller was compared in close to real-life power converter switching environments, demonstrating that the proposed RNN controller can achieve close to ideal optimal control performance even under low sampling rate conditions. The excellent performance of the proposed RNN controller under challenging and distorted system conditions further indicates the feasibility of using an RNN to approximate optimal control in practical applications.
Article
In recent years, deep neural networks (including recurrent ones) have won numerous contests in pattern recognition and machine learning. This historical survey compactly summarises relevant work, much of it from the previous millennium. Shallow and deep learners are distinguished by the depth of their credit assignment paths, which are chains of possibly learnable, causal links between actions and effects. I review deep supervised learning (also recapitulating the history of backpropagation), unsupervised learning, reinforcement learning & evolutionary computation, and indirect search for short programs encoding deep and large networks.
Article
Heterogeneous photocatalysis has been intensively studied in recent decades because it only requires photonic energy to activate the chemical conversion contrasting with conventional catalysis which needs heat for thermo-activation. Over the years, the theories for photochemical activity of photocatalyst including photo-induced redox reaction and super-hydrophilic conversion of TiO2 itself have been established. The progress in academic research significantly promotes its practical applications, including the field of photocatalytic construction and building materials. TiO2 modified building materials are most popular because TiO2 has been traditionally used as a white pigment. The major applications of TiO2 based photocatalytic building materials include environmental pollution remediation, self-cleaning and self-disinfecting. The advantage of using solar light and rainwater as driving force has opened a new domain for environmentally friendly building materials. In this paper, the basic reaction mechanisms on photocatalyst surface under the irradiation of ultraviolet and their corresponding applications in building and construction materials are reviewed. The problems faced in practical applications and the trends for future development are also discussed.
Article
The rates of the two main stages of photocatalytic ethanol destruction—oxidation of ethanol to acetaldehyde and oxidation of acetaldehyde to CO2—were studied under varied concentrations of ethanol and acetaldehyde and photocatalyst irradiance, at different temperatures, and over different photocatalysts. The rates followed the semiempirical three sites Langmuir–Hinshelwood model that envisages sites for ethanol and acetaldehyde adsorption and additional sites for competitive adsorption of ethanol and acetaldehyde. The increase in irradiance gave rise to higher selectivity toward CO2 via increased concentrations of gaseous intermediate acetaldehyde. However, the rate of ethanol oxidation rose faster than the rate of acetaldehyde oxidation. The selectivity toward CO2 monotonically decreased with temperature over TiO2 and the rate of oxidation reached a maximum at 80°C. Among platinum-doped catalysts, the best activity was found for 1.1% Pt/TiO2. Platinum addition to TiO2 resulted in a 1.5- to 2-fold increased overall rate of oxidation. The selectivity to CO2 over Pt/TiO2 catalyst monotonically increased with temperature. Separate studies in a batch reactor demonstrated that addition of platinum changed the product distribution. Acetic acid, instead of carbon monoxide, was formed in copious quantities over the Pt/TiO2 catalyst.
Article
The premise of this article is that learning procedures used to train artificial neural networks are inherently statistical techniques. It follows that statistical theory can provide considerable insight into the properties, advantages, and disadvantages of different network learning methods. We review concepts and analytical results from the literatures of mathematical statistics, econometrics, systems identification, and optimization theory relevant to the analysis of learning in artificial neural networks. Because of the considerable variety of available learning procedures and necessary limitations of space, we cannot provide a comprehensive treatment. Our focus is primarily on learning procedures for feedforward networks. However, many of the concepts and issues arising in this framework are also quite broadly relevant to other network learning paradigms. In addition to providing useful insights, the material reviewed here suggests some potentially useful new training methods for artificial neural networks.
Article
The authors propose a simulation tool (ST) able to test real-time hybrid GNSS/terrestrial and cooperative positioning algorithms that fuse both pseudorange measurements from satellites and terrestrial range measurements based on radio frequency communication performed between nodes of a wireless network. In particular, the ST simulates devices belonging to a peer-to-peer (P2P) wireless network where peers, equipped also with a GNSS receiver, cooperate among them by exchanging aiding data in order to improve both positioning accuracy and availability. Furthermore, the authors propose a method to increase the robustness of cooperative algorithms based on the estimated position covariance matrix. In particular, the proposed approach assures a faster estimation convergence and improved accuracy while lowering computational complexity and network traffic. Finally, the authors tested the sensitivity of the implemented positioning algorithms through the ST in two different scenarios, first in presence of high level of pseudorange noise and then in presence of a malicious peer in the P2P network.
Conference Paper
The backpropagation learning algorithm for neural networks is developed into a formalism for nonlinear signal processing. We illustrate the method by selecting two common topics in signal processing, prediction and system modelling, and show that nonlinear applications can be handled extremely well by using neural networks. The formalism is a natural, nonlinear extension of the linear Least Mean Squares algorithm commonly used in adaptive signal processing. Simulations are presented that document the additional performance achieved by using nonlinear neural networks. First, we demonstrate that the formalism may be used to predict points in a highly chaotic time series with orders of magnitude increase in accuracy over conventional methods including the Linear Predictive Method and the Gabor-Volterra-Weiner Polynomial Method. Deterministic chaos is thought to be involved in many physical situations including the onset of turbulence in fluids, chemical reactions and plasma physics. Secondly, we demonstrate the use of the formalism in nonlinear system modelling by providing a graphic example in which it is clear that the neural network has accurately modelled the nonlinear transfer function. It is interesting to note that the formalism provides explicit, analytic, global, approximations to the nonlinear maps underlying the various time series. Furthermore, the neural net seems to be extremely parsimonious in its requirements for data points from the time series. We show that the neural net is able to perform well because it globally approximates the relevant maps by performing a kind of generalized mode decomposition of the maps. 24 refs., 13 figs.
Article
Artificial neural networks now are used in many fields. They have become well established as viable, multipurpose, robust computational methodologies with solid theoretic support and with strong potential to be effective in any discipline, especially medicine. For example, neural networks can extract new medical information from raw data, build computer models that are useful for medical decision-making, and aid in the distribution of medical expertise. Because many important neural network applications currently are emerging, the authors have prepared this article to bring a clearer understanding of these biologically inspired computing paradigms to anyone interested in exploring their use in medicine. They discuss the historical development of neural networks and provide the basic operational mathematics for the popular multilayered perceptron. The authors also describe good training, validation, and testing techniques, and discuss measurements of performance and reliability, including the use of bootstrap methods to obtain confidence intervals. Because it is possible to predict outcomes for individual patients with a neural network, the authors discuss the paradigm shift that is taking place from previous “bin-model” approaches, in which patient outcome and management is assumed from the statistical groups in which the patient fits. The authors explain that with neural networks it is possible to mediate predictions for individual patients with prevalence and misclassification cost considerations using receiver operating characteristic methodology. The authors illustrate their findings with examples that include prostate carcinoma detection, coronary heart disease risk prediction, and medication dosing. The authors identify and discuss obstacles to success, including the need for expanded databases and the need to establish multidisciplinary teams. The authors believe that these obstacles can be overcome and that neural networks have a very important role in future medical decision support and the patient management systems employed in routine medical practice. Cancer 2001;91:1615–35. © 2001 American Cancer Society.
Article
This paper presents a literature review of using photocatalytic oxidation (PCO) to destruct volatile organic compounds (VOCs) in indoor air. TiO2 is used extensively as a photocatalyst due to its superior characteristics. Through kinetic experiments, the dependence of reaction rate on some key influencing factors (moisture, light intensity, initial contaminant concentration) has been studied, and kinetic models have been developed to aid the optimal reactor design. In general, the final products of PCO include CO2 and H2O. However, the intermediates, which are produced in the process of PCO, shouldn't be ignored because they can occupy the active sites of catalyst and lead to the deactivation of the catalyst.
Article
Volatile organic compounds (VOCs) are prevalent components of indoor air pollution. Among the approaches to remove VOCs from indoor air, photocatalytic oxidation (PCO) is regarded as a promising method. This paper is a review of the status of research on PCO purification of VOCs in indoor air. The review and discussion concentrate on the preparation and coating of various photocatalytic catalysts; different kinetic experiments and models; novel methods for measuring kinetic parameters; reaction pathways; intermediates generated by PCO; and an overview of various PCO reactors and their models described in the literature. Some recommendations are made for future work to evaluate the performance of photocatalytic catalysts, to reduce the generation of harmful intermediates and to design new PCO reactors with integrated UV source and reaction surface.
Article
The difficulty of Computational Wind Engineering (hereafter CWE) is described from the viewpoints of Computational Fluid Dynamics (hereafter CFD) technique. The rapid growth of CFD applications to wind engineering is presented. The new trends in turbulence models for applying CWE are noted. The advantages of dynamic subgrid scale (hereafter SGS) models in Large Eddy Simulation (hereafter LES) are clarified.
Article
This paper first discusses the aerodynamic effects of trees on local scale flow and pollutant concentration in idealized street canyon configurations by means of laboratory experiments and Computational Fluid Dynamics (CFD). These analyses are then used as a reference modelling study for the extension a the neighbourhood scale by investigating a real urban junction of a medium size city in southern Italy. A comparison with previous investigations shows that street-level concentrations crucially depend on the wind direction and street canyon aspect ratio W/H (with W and H the width and the height of buildings, respectively) rather than on tree crown porosity and stand density. It is usually assumed in the literature that larger concentrations are associated with perpendicular approaching wind. In this study, we demonstrate that while for tree-free street canyons under inclined wind directions the larger the aspect ratio the lower the street-level concentration, in presence of trees the expected reduction of street-level concentration with aspect ratio is less pronounced. Observations made for the idealized street canyons are re-interpreted in real case scenario focusing on the neighbourhood scale in proximity of a complex urban junction formed by street canyons of similar aspect ratios as those investigated in the laboratory. The aim is to show the combined influence of building morphology and vegetation on flow and dispersion and to assess the effect of vegetation on local concentration levels. To this aim, CFD simulations for two typical winter/spring days show that trees contribute to alter the local flow and act to trap pollutants. This preliminary study indicates that failing to account for the presence of vegetation, as typically practiced in most operational dispersion models, would result in non-negligible errors in the predictions.
Article
The improved computation presented in this paper is aimed to optimize the neural networks learning process using Levenberg-Marquardt (LM) algorithm. Quasi-Hessian matrix and gradient vector are computed directly, without Jacobian matrix multiplication and storage. The memory limitation problem for LM training is solved. Considering the symmetry of quasi-Hessian matrix, only elements in its upper/lower triangular array need to be calculated. Therefore, training speed is improved significantly, not only because of the smaller array stored in memory, but also the reduced operations in quasi-Hessian matrix calculation. The improved memory and time efficiencies are especially true for large sized patterns training.
Article
Artificial neural networks (ANNs) are biologically inspired computer programs designed to simulate the way in which the human brain processes information. ANNs gather their knowledge by detecting the patterns and relationships in data and learn (or are trained) through experience, not from programming. An ANN is formed from hundreds of single units, artificial neurons or processing elements (PE), connected with coefficients (weights), which constitute the neural structure and are organised in layers. The power of neural computations comes from connecting neurons in a network. Each PE has weighted inputs, transfer function and one output. The behavior of a neural network is determined by the transfer functions of its neurons, by the learning rule, and by the architecture itself. The weights are the adjustable parameters and, in that sense, a neural network is a parameterized system. The weighed sum of the inputs constitutes the activation of the neuron. The activation signal is passed through transfer function to produce a single output of the neuron. Transfer function introduces non-linearity to the network. During training, the inter-unit connections are optimized until the error in predictions is minimized and the network reaches the specified level of accuracy. Once the network is trained and tested it can be given new input information to predict the output. Many types of neural networks have been designed already and new ones are invented every week but all can be described by the transfer functions of their neurons, by the learning rule, and by the connection formula. ANN represents a promising modeling technique, especially for data sets having non-linear relationships which are frequently encountered in pharmaceutical processes. In terms of model specification, artificial neural networks require no knowledge of the data source but, since they often contain many weights that must be estimated, they require large training sets. In addition, ANNs can combine and incorporate both literature-based and experimental data to solve problems. The various applications of ANNs can be summarised into classification or pattern recognition, prediction and modeling. Supervised 'associating networks can be applied in pharmaceutical fields as an alternative to conventional response surface methodology. Unsupervised feature-extracting networks represent an alternative to principal component analysis. Non-adaptive unsupervised networks are able to reconstruct their patterns when presented with noisy samples and can be used for image recognition. The potential applications of ANN methodology in the pharmaceutical sciences range from interpretation of analytical data, drug and dosage form design through biopharmacy to clinical pharmacy.
Article
The health effects of air pollution have been subject to intense study in recent years. Exposure to pollutants such as airborne particulate matter and ozone has been associated with increases in mortality and hospital admissions due to respiratory and cardiovascular disease. These effects have been found in short-term studies, which relate day-to-day variations in air pollution and health, and long-term studies, which have followed cohorts of exposed individuals over time. Effects have been seen at very low levels of exposure, and it is unclear whether a threshold concentration exists for particulate matter and ozone below which no effects on health are likely. In this review, we discuss the evidence for adverse effects on health of selected air pollutants.
Photocatalytic application of titanium dioxide in architectural concrete: A review.
  • A.Srivastava