Dietrich Hartmann

Multiagent-Based Collaborative Framework for a Self-Managing Structural Health Monitoring System

The deterioration of civil infrastructure as a result of aging, altered requirements, excessive loading, or inadequate maintenance underpins the urgent need for reliable and cost-effective monitoring systems. This paper presents a framework for monitoring the condition of civil infrastructure. A self-managing software framework on the basis of multiagent technology is designed to remotely access and autonomously process collected information about the monitored structure. The distributed software framework ensures automated anomaly detection, supports collaborative diagnostic tools, and enhances communications among distributively located users participating in the monitoring activities. The multiagent framework has been implemented and validated for the monitoring of a 500-kW wind turbine in Germany. The long-term field instrumentation shows the practicability, efficiency, fault tolerance, and robustness of the system for structural health monitoring applications. This research has demonstrated a practic...

Comparison of Building Collapse Simulation Results From Finite Element and Rigid Body Models

In case of planning a building demolition, the information about geometry, quality of building materials, the design of the load carrying system and documentation of the structural calculation is often incomplete and imprecise. Thus for the analysis of a collapse event, engineers are forced to consider the uncertainty of primary parameters influencing e. g. the resistance of structural elements of a building. This kind of uncertainty can be described using suitable data models such as fuzziness and fuzzy randomness [6]. Within such an ‘uncertain’ structural analysis the deterministic fundamental solution is applied repeatedly. A comprehensive overview over algorithms of fuzzy analysis and fuzzy stochastic analysis is given in [5]. First applications of uncertainty collapse analyses can be found in [7, 8]. However, considering several uncertain parameters in an analysis the problem dimension and the necessary effort can be quite high. To receive a good prediction for a complex building collapse, several hundred or even more deterministic solutions are needed. This requires an efficient and fast scheme to perform the analysis for highly nonlinear problems, concerning geometry, material and changing boundary conditions such as contact.

Structural Health Monitoring based on Artificial Intelligence Techniques

Artificial Intelligence (AI) has a long history in computer science and is now being applied to engineering problems in Structural Health Monitoring (SHM) that would be difficult to solve by standard numerical techniques alone. In particular, the methods of Conventional Artificial Intelligence (CAI) and Computational Intelligence (CI), coupled with agent technology, show great promise in delivering monitoring systems that are robust, redundant, environmentally aware, economically sound as well as user friendly and highly adaptive. In this paper, background concepts of AI and an example of a SHM system for monitoring civil engineering structures are presented to clearly demonstrate the potential of intelligent software applications in the field of SHM.

A computational framework for life-cycle management of wind turbines incorporating structural health monitoring

The integration of structural health monitoring into life-cycle management strategies can help facilitating a reliable operation of wind turbines and reducing the life-cycle costs significantly. This article presents a life-cycle management framework for online monitoring and performance assessment of wind turbines, enabling optimum maintenance and inspection planning at minimum associated life-cycle costs. Incorporating continuously updated monitoring data (i.e. structural, environmental, and operational data), the framework allows capturing and understanding the actual wind turbine condition and, hence, reduces uncertainty in structural responses as well as load effects acting on the structure. As will be shown in this article, the framework integrates a variety of heterogeneous hardware and software components, including sensors and data acquisition units, server systems, Internet-enabled user interfaces as well as finite element models for system identification, and a multiagent system for self-detecting sensor malfunctions. To validate its capabilities and to demonstrate its practicability, the framework is deployed for continuous monitoring and life-cycle management of a 500-kW wind turbine. Remote life-cycle analyses of the monitored wind turbine are conducted, and case studies are presented investigating both the structural performance and the operational efficiency of the wind turbine.

A distributed agent-based approach for simulation-based optimization

Structural design and optimization in engineering are increasingly addressing non-standard optimization problems (NSPs). These problems are characterized by a complex topology of the optimization space with respect to nonlinearity, multimodality, discontinuity, etc. By that, NSP can only be solved by means of computer simulations. In addition, the corresponding numerical approaches applied often tend to be noisy. Typical examples for NSP occur in robust optimization, where the solution has to be robust with respect to implementation errors, production tolerances or uncertain environmental conditions. However, a generally applicable strategy for solving such problem categories always equally efficiently is not yet available. To improve the situation, a distributed agent-based optimization approach for solving NSPs is introduced in this paper. The elaborated approach consists of a network of cooperating but also competing strategy agents that wrap various strategies, especially optimization methods (e.g. SQP, DE, ES, PSO, etc.) using different search characteristics. In particular, the strategy agents contain an expert system modeling their specific behavior in an optimization environment by means of rules and facts on a highly abstract level. Further, different common interaction patterns have been defined to describe the structure of a strategy network and its interactions. For managing the complexity of NSPs using multi-agent systems (MASs) efficiently, a simulation and experimentation platform has been developed. Serving as a computational steering tool, it applies MAS technology and accesses a network of various optimization strategies. As a consequence, an elegant interactive steering, a customized modeling and a powerful visualization of structural optimization processes are established. To demonstrate the far reaching applicability of the proposed approach, numerical examples are discussed, including nonlinear function and robust optimization problems. The results of the numerical experiments illustrate the potential of the agent-based strategy network approach for collaborative solving, where observed synergy effects lead to an effective and efficient solution finding.

Optimization of slender structures considering geometrical imperfections

In this article, we present an optimization model which incorporates uncertainty induced by geometrical imperfections. Within the model, geometrical imperfections are represented by means of random fields. The induced uncertainties are then treated using the concept of a convex model. The resultant problem is then solved in a two-stage optimization procedure. An arched girder is used to demonstrate the capabilities of the proposed approach.

Implementation of a Multiagent-Based Paradigm for Decentralized Real-Time Structural Health Monitoring

Engineering structures such as wind turbines require continuous monitoring to ensure structural safety, to reduce the overall maintenance and repair costs and, ultimately, to achieve extended lifetimes and a greater economic viability. For that purpose, an automated SHM system for wind turbines has been developed and installed on a 500 kW wind turbine in Germany. During its operation, temporary malfunctions of the installed sensing units have been observed. These malfunctions, such as temporary sensor breakdowns which are well known from real-time SHM systems, might cause the loss of valuable monitoring data if not detected timely. A multi-agent system, which is capable of self-detecting system malfunctions and notifying the human individuals automatically, has been developed and integrated into the existing SHM system. The SHM system and its subsystem, the multi-agent system, have been in continuous operation since 2009. Since then, various malfunctions have automatically been detected and appropriate actions have been taken in a timely manner. As a result, the malfunctioning of the sensors did not lead to significant data loss, thus enhancing the quality of the SHM system.

An Object-Oriented Approach to High Order Finite Element Analysis of Three-Dimensional Continua

It has been shown recently that the p-version of the finite element method is well suited for the analysis of thin walled three-dimensional continua [2]. However, designing and implementing software for the p-version of FEM is a challenge because of the increased complexity compared to the h-version. In this paper, we present an object-oriented finite element system implemented in Java. It is pointed out how the object-oriented paradigm, suitable design patterns and thorough unit testing can help to develop and maintain a complex engineering application. The basic idea of the software design is to separate generally applicable mathematical concepts, like basis functions and geometrical mappings from concrete element formulations that contain the physics of the actual problem. In the mathematical package, there are interfaces representing the concepts of basis functions (forming an Ansatz space) and functions defined on R 1, R 2 and R 3. Several concrete classes implement these basic interfaces and realize for instance Lagrangian basis functions, hierarchical basis functions, functions constructed by a linear combination of basis functions or functions constructed by the blending function method. The use of NURBS curves hereby allows for the representation of complex geometries like that of the shell structure shown below. The power of this approach lies in the fact that on the element level only the interface types are used. Thus, an element formulation solely contains the physics and is not limited to a certain type of Ansatz space. Also, new Ansatz spaces can be easily incorporated lateron. Generally Java is considered as being slow for numerically intense applications. On the other side there are many advantages that make Java attractive also for simulation software [1]. In this paper, a hybrid approach is presented where the overall program is implemented in Java but numerically intense linear algebra operations are delegated to native code. In the full paper, it is shown that the software can be easily applied to problems involving 75 000 DOFs and more. Open image in new window

Multivariate Analysis and Prediction of Wind Turbine Response to Varying Wind Field Characteristics Based on Machine Learning

Site-specific wind field characteristics have a significant impact on the structural response and the lifespan of wind turbines. This paper presents a machine learning approach towards analyzing and predicting the response of wind turbine structures to varying wind field characteristics. Machine learning algorithms are applied (i) to better understand changes of wind field characteristics due to atmospheric conditions and (ii) to gain new insights into the wind turbine loads being affected by fluctuating wind. Using Gaussian Mixture Models, the variations in wind fields are investigated by comparing the joint probability distribution functions of several wind field features, which are constructed from long-term monitoring data taken from a 500 kW wind turbine in Germany, which is used as a reference system. Furthermore, based on Gaussian Discriminative Analysis, representative daytime and nocturnal wind turbine loads are predicted, compared, and analyzed.

Überblick zum Themenbereich Agentensysteme

Durch die im DFG Schwerpunktprogramm SPP 1103 durchgefuhrten Forschungsarbeiten zur agentenbasierten Software-Entwicklung wurde der Nachweis erbracht, dass die in der modernen Informatik entstandene Agententechnologie auch fur den Bereich der vernetzt-kooperativen Bauplanung hervorragend als Problemlosungsstrategie geeignet ist. Das Planen, das Entwerfen, das Konstruieren und die sich daran anschliesende Bauausfuhrung werden auf eine neue tragfahige Grundlage gestellt. Es ist klar geworden, dass die in der Bauplanung anstehenden grosen Zukunftsaufgaben, d. h. komplexe und durch viele Beteiligte gepragte Arbeitsprozesse bzw. Aktivitaten des Bauwesens fristgerecht, technisch und wirtschaftlich optimal abzuwickeln, durch die Agententechnologie — wie durch keine andere Technologie zuvor — adaquat unterstutzt werden. Es ist nunmehr moglich, die in der Wirklichkeit ablaufenden Arbeitsablaufe weitgehend frei von Informationsverlusten in ein Computermodell zu ubertragen; die Ubertragung erfolgt anschaulich sowie plausibel und kann jederzeit leicht nachgepruft werden. Dem normalen Sprachgebrauch folgend, wird dabei ein „Agent“ auch in der Informatik als „etwas“ angesehen, das — weitgehend selbststandig — im Auftrag oder im Interesse einer Person, einer Institution oder einer Organisation handelt.