Radislav Smid

Condition Indicators for Gearbox Condition Monitoring Systems

Condition monitoring systems for manual transmissions based on vibration diagnostics are widely applied in industry. The systems deal with various condition indicators, most of which are focused on a specific type of gearbox fault. Frequently used condition indicators (CIs) are described in this paper. The ability of a selected condition indicator to describe the degree of gearing wear was tested using vibration signals acquired during durability testing of manual transmission with helical gears.

Quality-Based Multiple-Sensor Fusion in an Industrial Wireless Sensor Network for MCM

The early alert monitoring system for an effective scheduled maintenance strategy based on a wireless technology requires reliable transfer of the diagnostic information between the sensor and the gateway. This paper presents an industrial wireless sensor network (IWSN)-based machine condition monitoring (MCM) system capable of overcoming a false indication caused by temporary loss of data, signal interference, or invalid data. We use multisensor fusion driven by a quality parameter, which is produced by each sensor node according to the data history outliers and the actual state of the node. The fusion node provides a quality evaluation on its output as well. This novel approach enables the propagation of information about the uncertainty of a measured value from the source node to the sink node. Thus, potential degradation of acquired or transferred diagnostic information is minimized. Instead of raw data, the signal features are transferred, so that bandwidth savings are rapidly improved. The proposed concept was experimentally verified on real wireless sensor network (WSN) hardware. The performance evaluated using the signal-to-noise ratio and false-alarm rate detection demonstrates the effectiveness of the proposed approach.

A Distributed Fault Detection System Based on IWSN for Machine Condition Monitoring

This paper introduces a novel framework for industrial wireless sensor networks (IWSNs) used for machine condition monitoring (MCM). Our approach enables the use of state-of-the-art computationally intensive classifiers in computationally weak sensor network nodes. The key idea is to split data acquisition, classifier building and training, and the operation phase, between different units. Computationally demanding processing is carried out in the central unit, while other tasks are distributed to the sensor nodes using over-the-air programming. The system is autonomously trained on the healthy state of a machine and then monitors a change in behavior which indicates a faulty state. Thanks to one-class classification, there is no need to introduce the faulty state of the machine in the training phase. We extend the diagnostic capability of the system using dynamic changes in the data acquisition and classification parts of the program in the sensor nodes. This enables the system to react to ambiguous machine states by temporarily changing the diagnostic focus. Compressing the information in the individual sensor nodes provided by in-node classification allows us to transmit only the classification result, instead of full signal waveforms. This enables the MCM system to be deployed with a large number of nodes, even with high sampling rates. The proposed concept was evaluated in IRIS IWSN by means of a rotary machine simulator.

Signal-to-noise ratio enhancement based on wavelet filtering in ultrasonic testing.

In ultrasonic non-destructive testing of materials with a coarse-grained structure the scattering from the grains causes backscattering noise, which masks flaw echoes in the measured signal. Several filtering methods have been proposed for improving the signal-to-noise ratio. In this paper we present a comparative study of methods based on the wavelet transform. Experiments with stationary, discrete and wavelet packet de-noising are evaluated by means of signal-to-noise ratio enhancement. Measured and simulated ultrasonic signals are used to verify the proposed de-noising methods. For comparison, we use signal-to-noise ratio enhancement related to fault echo amplitudes and filtering efficiency specific for ultrasonic signals. The best results in our setup were achieved with the wavelet packet de-noising method.

Development of a switched integrator amplifier for high-accuracy optical measurements

In the field of low flux optical measurements, the development and use of large area silicon detectors is becoming more frequent. The current/voltage conversion of their photocurrent presents a set of problems for traditional transimpedance amplifiers. The switched integration principle overcomes these limitations. We describe the development of a fully characterized current-voltage amplifier using the switched integrator technique. Two distinct systems have been developed in parallel at the United Kingdoms National Physical Laboratory (NPL) and Czech Metrology Institute (CMI) laboratories. We present the circuit theory and best practice in the design and construction of switched integrators. In conclusion the results achieved and future developments are discussed.

Classification of ultrasonic signals

In ultrasonic non-destructive testing it is very difficult to detect flaws in materials with coarse-grain structure. The ultrasonic signals measured on these materials contain echoes which are very similar to fault echoes. These echoes arise from grains which are contained in the material. For the detection of flaws various methods for suppressing echoes from grains have to be used. In this work we used the method for filtering ultrasonic signals based on discrete wavelet transform. For the classification of ultrasonic signals in A-scan we used a pattern recognition method called support vector machines. In this study we classify signals with fault echoes, echo from weld and back-wall echo. Ultrasonic signals were measured on materials used for constructing aeroplane engines. The experimental results indicate the performance of the proposed approach.

Thickness measurement using transient eddy current techniques

This paper describes a pulsed eddy current (PEC) based non-destructive testing system used for experimental evaluation of material thickness and properties. Both classical and modern sensing methods of magnetic field have been carried out. An inductive as well as a giant magnetoresistive sensor based (GMR) probe have been used for sensing and the results obtained are presented and compared. The use of GMRs for this type of application is innovative and the post processing method too, since it evaluates the thicknesses of the conductive samples or of its non-conductive coatings, from the values obtained with the integration of measured signals.

Signal-to-Noise Ratio Improvement based on the Discrete Wavelet Transform in Ultrasonic Defectoscopy

In ultrasonic testing it is very important to recognize the fault echoes buried in a noisy signal. The fault echo characterizes a flaw in the material. An important requirement on ultrasonic signal filtering is zero-time shift, because the position of ultrasonic echoes is essential. This requirement is accomplished using the discrete wavelet transform (DWT), which is used for reducing the signal-to-noise ratio. This paper evaluates the quality of filtering using the discrete wavelet transform. Additional computer simulations of the proposed algorithms are presented.

Various scale errors in dithered quantizers: Visualisation and reduction

Abstract Dithering is being used for elimination quantization error and other small scale errors of A/D converters. However, large scale errors remain. It is possible to reduce these errors by the look-up table or by subtracting a suitable approximation (e.g. polynomian or spline). Sub-quantum accuracy can be achieved. The prediction of metrological reliability of the dithered AD converter cannot be simply made. It is necessary to separate errors of various scales of the used AD converter. The multiscale approach has been used for several years for characterization of complex signals (like ECG, seismic signals etc.). One of the multiscale analysis methods, the Continuous Wavelet Transform (CWT), makes it possible to find out at which scale and position the particular phenomenon occurred.

Reducing the calibration points of multisensors

Goal of this paper is to develop the method identification of multisensors characteristic curve using neural networks and minimize the number of calibration points required to find an appropriate characteristic curve for the input-to-output relationship of multisensors. A proposed original method illustrates how to improve the prediction accuracy of calibration results for multisensors despite the high level of random error.