With increasing complexity of systems it is In this paper, fast to develop and easy to implement becoming more and more time consuming and difficult fault detection methods are proposed for different systems. to achieve reliable fault detection strategies. Using model-The approach is compared to classical signal-based fault based methods requires detailed knowledge about the systems behavior and seems in some cases successful in theory but undetectintrateg in fixed theshlds.iThe ain id applicable in real-time due to high computation requirements. is briefly introduced in section II, the classical approach isIn this contribution, an idea and algorithm for feature-based described in section III. It will be shown that both methods fault detection approach is proposed. The main idea of this are satisfyingly detecting faults but with the difference that approach is to detect and identify faults in a complex system the faults can be identified using the proposed method. A without any kind of modeling. By extracting features from relevant sensor signals, yielded from Hardware-in-the-Loop way to realize it by using quantized data in a matrix iS simulations, and combining them in a matrix, it is possible for proposed in section IV. Experimental results and compara human operator to denote subsets of the matrix as fault-free isons are presented in section V and conclusions are given and faulty areas. An advantage is the ability to set individual in section VI.thresholds for the subsets, giving a more robustness towards false alarms and a possibility to denote individual subsets to II. IDEA relevant faults. From this, it will be shown that identification of faults is possible. In order to achieve a fault detectionThe increasing demand of time gain from ideas to and identification ability, it is necessary to implement the research and implementation of software solutions in mass faults of interest in a test rig and conduct Hardware-in-theproduction components makes it necessary to adapt also Loop simulations. The raw data from fault-free and faulty FDI demands to the market restrictions. Lowering costs are simulations are used in the training of the matrix and the algorithm detects and identifies the faults in a robust way. Th essental for today's manufacturers n different areas, e.g.results are compared to a classical fault detection method that industry, robotics, and vehicle dynamics systems. In the uses fixed thresholds. It will be shown how a sensor bias fault field of FDI many methods have been developed to suit and a pressure relief valve fault are detected and identified. different purposes. Terms like model-free, adapted, datadriven, and signal-based are gaining more ground [3]. The I. INTRODUCTION main idea of the approach introduced in this paper is to As systems become more complex the sensitivity for use the human knowledge and pragmatism to develop fast, faults increases. The increasing demand from customers reliable, and time efficient (in sense of development time) makes it necessary to develop robust and a...
