Refine
Document Type
- Conference Proceeding (16)
- Article (1)
Institute
Keywords
- Computerunterstütztes Verfahren (9)
- Architektur <Informatik> (6)
- Angewandte Informatik (4)
- Angewandte Mathematik (4)
- CAD (4)
- Verteiltes System (3)
- Bauwerk (2)
- Computer Science Models in Engineering; Multiscale and Multiphysical Models; Scientific Computing (2)
- Finite-Elemente-Methode (2)
- Ingenieurbau (2)
Ziel des Teilprojekts C5 des Sonderforschungsbereichs 398 an der Ruhr-Universität Bochum ist die Entwicklung eines lebensdauerorientierten Entwurfsmodells für Stahltragwerke. Das Referenzsystem der letzten Förderperiode war eine einschiffige Stahlhalle mit Kranbahnen. Der vorliegende Beitrag stellt die Ergebnisse einer Sensitivitätsanalyse des Stahlrahmens unter Kranfahrtlasten vor. Diese Verkehrslasten wurden, völlig entkoppelt von weiteren äußeren Einflüssen wie Wind und Schnee, als poissongetriebene Pulsprozesse beschrieben, die zufälligen Eigenschaften der Kranfahrtlasten wurden in den Zufallsvariablen Pulsdauer und Pulsintensität berücksichtigt. Zur Minimierung des Rechenaufwandes der Lebensdauer-untersuchung wurden a priori einige Systemparameter anhand von Parameterstudien im Hinblick auf ihre dynamische Wirkung maßgebend festgelegt. In einer darauf aufbauenden Betriebsfestigkeitsuntersuchung wurden die Schädigungseinträge in einer Rahmenecke infolge von jeweils einzelnen Pulslastereignissen berechnet und in einer Datenbasis abgelegt. Eine abschließende Distance-Controlled Monte Carlo Simulation der Pulsprozesse über einen maximalen Zeitraum von 100 Jahren überführte die Realisationen der einzelnen Pulslasten anhand der Datenbasis in Teilschädigungen, welche nach der Palmgren/Miner-Hypothese zu einer Gesamtschädigung aufsummiert wurden. Der Einfluss der Kranlasten auf den Entwurf von Stahlhallen wurde durch den Vergleich der berechneten Lebensdauer und der planmäßig vorgegebenen Nutzungsdauer quantifiziert.
In this contribution the software design and implementation of an analysis server for the computation of failure probabilities in structural engineering is presented. The structures considered are described in terms of an equivalent Finite Element model, the stochastic properties, like e.g. the scatter of the material behavior or the incoming load, are represented using suitable random variables. Within the software framework, a Client-Server-Architecture has been implemented, employing the middleware CORBA for the communication between the distributed modules. The analysis server offers the possibility to compute failure probabilities for stochastically defined structures. Therefore, several different approximation (FORM, SORM) and simulation methods (Monte Carlo Simulation and Importance Sampling) have been implemented. This paper closes in showing several examples computed on the analysis server.
Due to increasing numbers of wind energy converters, the accurate assessment of the lifespan of their structural parts and the entire converter system is becoming more and more paramount. Lifespan-oriented design, inspections and remedial maintenance are challenging because of their complex dynamic behavior. Wind energy converters are subjected to stochastic turbulent wind loading causing corresponding stochastic structural response and vibrations associated with an extreme number of stress cycles (up to 109 according to the rotation of the blades). Currently, wind energy converters are constructed for a service life of about 20 years. However, this estimation is more or less made by rule of thumb and not backed by profound scientific analyses or accurate simulations. By contrast, modern structural health monitoring systems allow an improved identification of deteriorations and, thereupon, to drastically advance the lifespan assessment of wind energy converters. In particular, monitoring systems based on artificial intelligence techniques represent a promising approach towards cost-efficient and reliable real-time monitoring. Therefore, an innovative real-time structural health monitoring concept based on software agents is introduced in this contribution. For a short time, this concept is also turned into a real-world monitoring system developed in a DFG joint research project in the authors’ institute at the Ruhr-University Bochum. In this paper, primarily the agent-based development, implementation and application of the monitoring system is addressed, focusing on the real-time monitoring tasks in the deserved detail.
In order to model and simulate collapses of large scale complex structures, a user-friendly and high performance software system is essential. Because a large number of simulation experiments have to be performed, therefore, next to an appropriate simulation model and high performance computing, efficient interactive control and visualization capabilities of model parameters and simulation results are crucial. To this respect, this contribution is concerned with advancements of the software system CADCE (Computer Aided Demolition using Controlled Explosives) that is extended under particular consideration of computational steering concepts. Thereby, focus is placed on problems and solutions for the collapse simulation of real world large scale complex structures. The simulation model applied is based on a multilevel approach embedding finite element models on a local as well as a near field length scale, and multibody models on a global scale. Within the global level simulation, relevant effects of the local and the near field scale, such as fracture and failure processes of the reinforced concrete parts, are approximated by means of tailor-made multibody subsystems. These subsystems employ force elements representing nonlinear material characteristics in terms of force/displacement relationships that, in advance, are determined by finite element analysis. In particular, enhancements concerning the efficiency of the multibody model and improvements of the user interaction are presented that are crucial for the capability of the computational steering. Some scenarios of collapse simulations of real world large scale structures demonstrate the implementation of the above mentioned approaches within the computational steering.
Die Bauwerksüberwachung gewinnt aus sicherheitstechnischen sowie aus wirtschaftlichen Gründen zunehmend an Bedeutung. Nicht nur die Bauwerkssicherheit kann durch leistungsfähige Monitoring-Systeme angemessen beurteilt, auch die Nutzungsdauer bestehender Bauwerke kann durch die gewonnenen Informationen deutlich verlängert werden. Das vorliegende Papier beschreibt die Entwicklung eines webbasierten Talsperren-Monitoring-Systems, das die automatisierte Erfassung von Daten vor Ort sowie die computergestützte Aufbereitung und Analyse der gesammelten Messdaten ermöglicht. Das Monitoring-System ist durch seinen modularen Aufbau nicht auf die Talsperren-Überwachung beschränkt, sondern kann ohne großen Aufwand an andere Überwachungsaufgaben angepasst werden. Das System besteht aus drei wesentlichen Modulen: (i) einer erweiterbaren Klassenbibliothek, die die Steuerung der im Bauwerk installierten Messelektronik ermöglicht, (ii) einem webbasierten Datenerfassungsmodul, das neben der automatischen Datenerfassung eine Fernsteuerung der Messelektronik erlaubt und Funktionen zur Verwaltung der Überwachungsaufgaben bereitstellt, sowie (iii) einem webbasierten Visualisierungs- und Auswertungsmodul zur Aufbereitung und Analyse der gesammelten Daten. Alle an der Überwachung beteiligten Mitarbeiter können mit einem üblichen Web-Browser über das Internet auf das entwickelte System zugreifen; ein Zugriff mittels Mobiltelefon ist alternativ möglich. Das implementierte Talsperren-Monitoring-System begleitet die beteiligten Fachleute von der Erfassung der Daten vor Ort bis hin zur Aufbereitung und Analyse der Messdaten an zentraler Stelle: Die Mitarbeiter werden durch einen einfachen Zugriff auf die installierte Messelektronik, automatisierte Messungen und umfangreiche Analysefunktionalitäten bei ihren spezifischen Aufgaben unterstützt. Der bisherige manuelle Arbeitsaufwand für Datenerfassung, -transfer und Analyse wird somit deutlich reduziert.
Due to the complex interactions between the ground, the driving machine, the lining tube and the built environment, the accurate assignment of in-situ system parameters for numerical simulation in mechanized tunneling is always subject to tremendous difficulties. However, the more accurate these parameters are, the more applicable the responses gained from computations will be. In particular, if the entire length of the tunnel lining is examined, then, the appropriate selection of various kinds of ground parameters is accountable for the success of a tunnel project and, more importantly, will prevent potential casualties. In this context, methods of system identification for the adaptation of numerical simulation of ground models are presented. Hereby, both deterministic and probabilistic approaches are considered for typical scenarios representing notable variations or changes in the ground model.
This paper presents a robust model updating strategy for system identification of wind turbines. To control the updating parameters and to avoid ill-conditioning, the global sensitivity analysis using the elementary effects method is conducted. The formulation of the objective function is based on M¨uller-Slany’s strategy for multi-criteria functions. As a simulationbased optimization, a simulation adapter is developed to interface the simulation software ANSYS and the locally developed optimization software MOPACK. Model updating is firstly tested on the beam model of the rotor blade. The defect between the numerical model and the reference has been markedly reduced by the process of model updating. The effect of model updating becomes more pronounced in the comparison of the measured and the numerical properties of the wind turbine model. The deviations of the frequencies of the updated model are rather small. The complete comparison including the free vibration modes by the modal assurance criteria shows the excellent coincidence of the modal parameters of the updated model with the ones from the measurements. By successful implementation of the model validation via model updating, the applicability and effectiveness of the solution concept has been demonstrated.
Interval analysis extends the concept of computing with real numbers to computing with real intervals. As a consequence, some interesting properties appear, such as the delivery of guaranteed results or confirmed global values. The former property is given in the sense that unknown numerical values are in known to lie in a computed interval. The latter property states that the global minimum value, for example, of a given function is also known to be contained in a interval (or a finite set of intervals). Depending upon the amount computation effort invested in the calculation, we can often find tight bounds on these enclosing intervals. The downside of interval analysis, however, is the mathematically correct, but often very pessimistic size of the interval result. This is in particularly due to the so-called dependency effect, where a single variable is used multiple times in one calculation. Applying interval analysis to structural analysis problems, the dependency has a great influence on the quality of numerical results. In this paper, a brief background of interval analysis is presented and shown how it can be applied to the solution of structural analysis problems. A discussion of possible improvements as well as an outlook to parallel computing is also given.
Although there are some good reasons to design engineering software as a stand-alone application for a single computer, there are also numerous possibilities for creating distributed engineering applications, in particular using the Internet. This paper presents some typical scenarios how engineering applications can benefit from including network capabilities. Also, some examples of Internet-based engineering applications are discussed to show how the concepts presented can be implemented.
This paper deals with two different agent-based approaches aimed at the incorporation of complex design information into multi-agent planning systems. The first system facilitates collaborative structural design processes, the second one supports fire engineering in buildings. Both approaches are part of two different research projects that belong to the DFG1 priority program 1103 entitled “Network-based Co-operative Planning Processes in Structural Engineering“ (DFG 2000). The two approaches provide similar database wrapper agents to integrate relevant design information into two multi-agent systems: Database wrapper agents make the relevant product model data usable for further agents in the multi-agent system, independent on their physical location. Thus, database wrapper agents act as an interface between multi-agent system and heterogeneous database systems. The communication between the database wrapper agents and other requesting agents presumes a common vocabulary: a specific database ontology that maps database related message contents into database objects. Hereby, the software-wrapping technology enables the various design experts to plug in existing database systems and data resources into a specific multi-agent system easily. As a consequence, dynamic changes in the design information of large collaborative engineering projects are adequately supported. The flexible architecture of the database wrapper agent concept is demonstrated by the integration of an XML and a relational database system.