Oliver Niggemann

On the Nature of Structure and Its Identification

When working on systems of the real world, abstractions in the form of graphs have proven a superior modeling and representation approach. This paper is on the analysis of such graphs. Based on the paradigm that a graph of a system contains information about the systems structure, the paper contributes within the following respects: 1. It introduces a new and lucid structure measure, the so-called weighted partial connectivity, Λ, whose maximization defines a graphs structure (Section 2). 2. It presents a fast algorithm that approximates a graphs optimum Λ-value (Section 3). Moreover, the proposed structure definition is compared to existing clustering approaches (Section 4), resulting in a new splitting theorem concerning the well-known minimum cut splitting measure. A key concept of the proposed structure definition is its implicit determination of an optimum number of clusters. Different applications, which illustrate the usability of the measure and the algorithm, round off the paper (Section 5).

Identifying behavior models for process plants

The increasing complexity of todays production systems and the variety of model-based approaches to their monitoring, diagnosis and testing emphasize the importance of the modeling step. Modeling is mostly done manually, in a costly and time-consuming way. In this paper, an alternative that comes from the learning theory is given: an automated procedure for identifying behavior models from recorded observations. Assuming the systems structure is known, the algorithm presented here is capable of learning behavior models for its components. The algorithm accounts for probabilistic, timing, discrete and continuous aspects of the given system, using the modeling formalism of hybrid automata. The practical usability of identified models is demonstrated using an anomaly detection application for a real production system.

Using behavior models for anomaly detection in hybrid systems

The importance of safety and reliability in todays real-world complex hybrid systems, such as process plants, led to the development of various anomaly detection and diagnosis techniques. Model-based approaches established themselves among the most successful ones in the field. However, they depend on a model of a system, which usually needs to be derived manually. Manual modeling requires a lot of efforts and resources. This paper gives a procedure for anomaly detection in hybrid systems that uses automatically generated behavior models. The model is learned from logged systems measurements in a hybrid automaton framework. The presented anomaly detection algorithm utilizes the model to predict the system behavior, and to compare it with the observed behavior in an online manner. Alarms are raised whenever a discrepancy is found between these two. The effectiveness of this approach is demonstrated in detecting several types of anomalies in a real-world running production system.

A stochastic method for the detection of anomalous energy consumption in hybrid industrial systems

In the presented work, the detection of anomalous energy consumption in hybrid industrial production systems is investigated. A model-based approach with a timed hybrid automaton as overall system model is employed for anomaly detection. The approach is based on the assumption of several system modes, i.e. phases with continuous system behavior. Transitions between the modes are attributed to discrete control events such as on/off signals. The underlying discrete event system which comprises both system modes and transitions is modeled as finite state machine. The focus of this paper is set on the modeling of the energy consumption in the particular system modes. Sequences of stochastic state space models are employed for this purpose. Model learning and anomaly detection for this approach are considered. The proposed approach is further evaluated in a small model factory. The experimental results show significant improvements compared to existing approaches to anomaly detection in hybrid industrial systems.

AutomationML: From data exchange to system planning and simulation

The planning, testing and integration of modern automation systems is becoming more and more a bottleneck in the construction of new production facilities. This is due to the facts that plants grow in complexity and that modern automation systems are highly distributed and comprise complex components. To cope with these challenges and to guarantee short implementation times and a small number of errors for the automation systems, modern development processes are needed. Such modern processes can be reduced to four main aspects: (i) A seamless process with corresponding seamless tools, (ii) a high level of model reuse and adaptability, (iii) executable models and early tests, and (iv) a system-wide planing process of the distributed system. Therefore, the established tool landscape with its specialized tools for each discipline of engineering has difficulties to keep up with these trends. The approach presented in this paper implements a development process including the aspects (i) (iv) using the new data exchange format AutomationML. AutomationML serves as an enabling technology and has the potential to change future development processes and may trigger the development of new, better integrated tools.

Models for model's sake: why explicit system models are also an end to themselves.

In automotive software and system design, explicit system and especially software models have found their way into the development process. This paper try to give an overview for what such models have so-far been used and which advantages they brought to vehicle manufacturers and suppliers. Another focus of this paper is the comparison to functional models which are already used in the automotive industry to define control algorithms and function implementation. In many cases too strong analogies have been seen between the existing functional control algorithm models and the new system models - leading to suboptimal development processes and tools. This paper therefore try to outline differences between these model types. Finally, a synthesis between functional, system, and software models was sketched.

Detecting anomalous energy consumptions in distributed manufacturing systems

This paper presents a novel model-based approach for the prediction of energy consumption in production plants in order to detect anomalies. A special Ethernet-based data acquisition approach is implemented that features real-time sampling of process and energy data. Hybrid timed automaton models of the supervised production plant are generated and executed in parallel to the system by using data samples as model input. According to comparisons of predicted energy consumption with the production plant observations, anomalies can be detected automatically. An evaluation within a small factory shows that anomalies of 10 % differences in energy consumption, wrong control sequences and wrong timings can be detected with a minimum accuracy of 98 %. With this approach, downtimes of production systems can be shortened and atypical energy consumptions can be detected and adjusted to optimal operation.

System modeling based on machine learning for anomaly detection and predictive maintenance in industrial plants

Electricity, water or air are some Industrial energy carriers which are struggling under the prices of primary energy carriers. The European Union for example used more 20.000.000 GWh electricity in 2011 based on the IEA Report [1]. Cyber Physical Production Systems (CPPS) are able to reduce this amount, but they also help to increase the efficiency of machines above expectations which results in a more cost efficient production. Especially in the field of improving industrial plants, one of the challenges is the implementation of anomaly detection systems. For example as wear-level detection, which improves maintenance cycles and thus leads to a better energy usage. This paper presents an approach that uses timed hybrid automata of the machines normal behavior for a predictive maintenance of industrial plants. This hybrid model reduces discrete and continuous signals (e.g. energy data) to individual states, which refer to either the present condition of the machines. This allows an effective anomaly detection by implementing a combined data acquisition and anomaly detection approach, and the outlook for other applications, such as a predictive maintenance planning. Finally, this methodology is verified by three different industrial applications.

A meta heuristic for graph drawing: learning the optimal graph-drawing method for clustered graphs

The problem of finding a pleasant layout for a given graph is a key challenge in the field of information visualization. For graphs that are biased towards a particular property such as tree-like, star-like, or bipartite, a layout algorithm can produce excellent layouts—if this property is actually detected. Typically, a graph may not be of such a homogeneous shape but is comprised of different parts, or it provides several levels of abstraction each of which dominated by another property. The paper in hand addresses the layout of such graphs. It presents a meta heuristic for graph drawing, which is based on two ideas: (i) The detection and exploitation of hierarchical cluster information to unveil a graphs inherent structure. (ii) The automatic selection of an individual graph drawing method for each cluster.

Energy efficiency optimization by automatic coordination of motor speeds in conveying systems

This paper addresses the optimization of energy flows between electric drives in conveying systems. Thereby, load peaks and the feedback of electric energy into the grid are reduced. The approach is based on the solution of a mixed integer quadratic optimization problem which incorporates models for energy flows and energy consumption of the electric drives. Models for energy flows and energy consumption can be parametrized from sensor data. Besides, movement constraints such as start positions, end positions and time limits are taken into account. The solution of the optimization problem is accomplished in realtime by application of standard solvers. Experimental results show that the proposed methods allows recovering regenerative energy of electric drives as motoric power for other drives. Integration into material flow planning of an automated warehouse is straightforward so that an inexpensive and simply usuable way for power saving in intralogistics is presented.