Standard bearing fault detection features are shown to be ineffective for estimating bearings' remaining useful life (RUL). Addressing this issue, we propose an approach for bearing fault prognostics, which employs Rényi entropy based features to describe the statistical properties of the envelope of the generated vibrations and a set of Gaussian process (GP) models to relate the feature value to RUL. GP models are non-parametric black-box models which search for the relationships among measured data rather than trying to approximate the modelled system by fitting the parameters of the selected basis functions. Bearing's RUL is estimated as a posterior distribution following the Bayes' rule using GP models' output as likelihood distribution. The proposed approach was evaluated on the data set provided for the IEEE PHM 2012 Prognostic Data Challenge.
E.02 International awards
COBISS.SI-ID: 25958695Oscillations in the control systems appear due to disturbances or inappropriate controller structure or parameters. Oscillations in the system result in increased power consumption, decrease of product quality and lower actuator’s life expectancy, which leads to higher production costs and stalls in production. Department of systems and control research group, together with the R/D group of the Danfoss Trata, d.o.o., developed innovative iMVCV motor drives for valves which reduce oscillations in control systems. The immediate impact of new product reflects in increased sales on the overall comopany level. Our group, together with the R/D group of the Danfoss Trata, d.o.o., was awarded state's Puch prize for innovative solution.
E.01 National awards
COBISS.SI-ID: 26053415The first prototype of a diagnostic and prognostic platform builds on distributed sensor network. Several sensors are connected to the so-called smart node which performs data acquisition within the prescribed acquisition sessions, does feature extraction and transmit the results to the central server. Various communication protocols are supported.
F.08 Development and manufacture of a prototype
COBISS.SI-ID: 25932327In last years in the area of tools and methods for control software design a model-driven development approach MAGICS (Modeling and Automatic Generation of Industrial process Control Software) was developed. In the year 2012 the approach was upgraded into a development methodology, which includes an enlarged set of activities and guidelines at two engineering levels, formalized modeling language ProcGraph, and supporting software tools. A new version of the ProcGraph graphic editor was developed, now including low level textual part of the models, which enables complete modeling of the software within the development environment. An empirical evaluation of the MAGICS methodology was performed in the form of a case study with participation of the developers from an industrial partner. The results of the evaluation showed that MAGICS, in relation to the actual approach, improves productivity by between 18% and 33% of total development effort and significantly improves software quality.
F.06 Development of a new product
COBISS.SI-ID: 265150464The diagnostic system for vacuum cleaner motors (type 495) is designed to perform end quality assessment of the vacuum motors at Domel, d.d. The system is positioned as the final station on a motor production line and is responsible for detection of any deviations of the produced motor units. The system measures electric parameters (voltage, current, power), as well as vibrations and units suction pressure, which are fed to diagnostic algorithms that determine whether the unit has passed the test. Additionally, the system analyses the commutation quality of the produced motors with purpose built diagnostic tool, which has builtin innovative algorithm for commutation quality assessment developed at our department. The diagnostic system is connected with factory's information center via Ethernet and can store diagnostic results to central repository for offline analysis. This enables optimal traceability of each produced unit and represents valuable feedback information for assessing the state of the production line as a whole.
F.09 Development of a new technological process or technology
COBISS.SI-ID: 26009895