J. Martinez-Roman

Harmonic Order Tracking Analysis: A Novel Method for Fault Diagnosis in Induction Machines

The diagnosis of induction machines using Fourier transform relies on tracking the frequency signature of each type of fault in the currents spectrum, but this signature depends on the machines slip and the supply frequency, so it must be recomputed for each working condition by trained personnel or by diagnostic software. Besides, sampling the current at high rates during long times is needed to achieve a good spectral resolution, which requires large memory space to store and process the current spectra. In this paper, a novel approach is proposed to solve both problems. It is based on the fact that each type of fault generates a series of harmonics in the currents spectrum, whose frequencies are multiples of a characteristic main fault frequency. The tracking analysis of the fault components using the harmonic order (defined as the frequency in per unit of the main fault frequency) as independent variable instead of the frequency generates a unique fault signature, which is the same for any working condition. Besides, this signature can be concentrated in just a very small set of values, the amplitudes of the components with integer harmonic order. This new approach is introduced theoretically and validated experimentally.

Low-Cost Diagnosis of Rotor Asymmetries in Induction Machines Working at a Very Low Slip Using the Reduced Envelope of the Stator Current

Fault diagnosis of rotor asymmetries in induction machines working at a very low slip, through Fourier-based methods, usually requires a long acquisition time to achieve a high spectral resolution and a high sampling frequency to reduce aliasing effects. However, this approach generates a huge amount of data, which makes its implementation difficult using embedded devices with small internal memory, such as digital signal processors and field programmable gate arrays or devices with low computing power. In this paper, a new simplified diagnostic signal designated as the reduced envelope of the stator current is introduced to address this problem. The reduced envelope signal is built using only one sample of the current per cycle without any further processing, and it is demonstrated that it carries the same spectral information about the fault as the full-length current signal. Based on this approach, an embedded device has only to store and process a minimal set of samples compared with the raw current signal for a desired resolution. In this paper, the theoretical basis of the proposed method is presented, as well as its experimental validation using two different motors with broken bars: 1) a high-power induction motor working in a factory; and 2) a low-power induction motor mounted in a laboratory test bed.

Short-Frequency Fourier Transform for Fault Diagnosis of Induction Machines Working in Transient Regime

Transient-based methods for fault diagnosis of induction machines (IMs) are attracting a rising interest, due to their reliability and ability to adapt to a wide range of IM’s working conditions. These methods compute the time–frequency (TF) distribution of the stator current, where the patterns of the related fault components can be detected. A significant amount of recent proposals in this field have focused on improving the resolution of the TF distributions, allowing a better discrimination and identification of fault harmonic components. Nevertheless, as the resolution improves, computational requirements (power computing and memory) greatly increase, restricting its implementation in low-cost devices for performing on-line fault diagnosis. To address these drawbacks, in this paper, the use of the short-frequency Fourier transform (SFFT) for fault diagnosis of induction machines working under transient regimes is proposed. The SFFT not only keeps the resolution of traditional techniques, such as the short-time Fourier transform, but also achieves a drastic reduction of computing time and memory resources, making this proposal suitable for on-line fault diagnosis. This method is theoretically introduced and experimentally validated using a laboratory test bench.

A Unified Approach to the Very Fast Torque Control Methods for DC and AC Machines

The general strategy to get a very fast torque control in a DC or AC machine is based on keeping the pulsational electromotive forces (EMFs) of all of its phases as small as possible during the transient states and achieving the torque changes by exclusively enhancing the rotational EMFs. All the resources available have to be oriented in this direction. This very simple but profound physical idea, when applied to DC or AC machines, allows the different control methods for these machines that have been developed to date and regarded as the best from a dynamic point of view to be deduced in a unified, systematic, and straightforward way.

Harmonic order tracking analysis: A speed-sensorless method for condition monitoring of wound rotor induction generators in wind turbines

This paper introduces a speed-sensorless method for detecting rotor asymmetries in wound rotor induction machine (WRIM) working as generator in wind turbines. The method is based on time-frequency analysis of rotors currents and a new transformation, which make it suitable for every working condition of the machine: steady state, transient regime and non-stationary conditions. Besides, the results are shown in the same sort of plot regardless the working conditions, reducing the number of point to be stored and highlighting the presence or absence of faults in a simple and clear way. Simulation and experimental results show the validity of the method to detect rotor electrical faults in WRIM under any working condition.

Condition monitoring of electrical machines using low computing power devices

The diagnosis of electrical motors through the detection of fault frequency signatures in the currents spectrum has become an established standard in the field of industrial maintenance systems. Nevertheles, its implementation on devices with low computing power remains a practical challenge. Industrial controllers, such as programmable logic controllers, or modern, low cost controller hardware, such as the Arduino or the Raspberry Pi open source hardware proposals, lack both the on-board memory and the high speed data acquisition hardware to perform an accurate spectral analysis of the machines current, in order to identify the spectral components produced by each type of fault. In this paper, a signal conditioning unit, based on a novel downsampling method of the current, is presented. This unit reduces the amount of current samples that must be processed by the diagnostic unit to a mere sample per current cycle, maintains the sub-hertz accuracy needed to resolve fault, and converts the mains component into a constant value that can be easily eliminated without using any additional filter. Besides, it is implemented using low cost devices, just resistors and operational amplifiers. The proposed method is theoretically developed in this paper, and it has been validated using induction motors with broken bars fed directly by the mains or through variable speed drives.

Multilayer Park's vector approach, a method for fault detection on induction motors

This paper introduces a new method of fault diagnosis for induction motors, the Multilayer Parks Vector Approach, derived from the well-known fault diagnostic method based on Parks Vector Approach analysis. This new method allows the detection of failures in steady and transient regimes operation on induction motors. The resulting vector generated provides the information about deviations in the circularity of Parks vector for each motor cycle, allowing with this information the diagnosis of failures. As advantages, this method have a lower computational cost than others methods derived from the fault analysis with Parks Vector Approach. This requires only knowledge of supply frequency and control method, without setting parameters and without motor specification. To test the reliability of this method, a test of broken rotor bars through Multilayer Parks Vector Approach has been done, obtaining relevant characteristics for the successful broken rotor bar detection.

Automatic translation of Programmable Logic Controllers (PLC) control programs in packaging machinery

Programmable Logic Controllers (PLC) are key components in modern packaging automation systems. But their success has led to the presence of a broad range of brands and models, each one with its own and unique features, but also with its own environment for software development. This situation has some major drawbacks: it is very difficult to upgrade an installation by changing the PLC, maintenance costs rise with the multiplicity of PLC brands in a given factory, and production parameters must be changed when the production is changed to a machine with a different PLC. In this paper, a first step is presented to address such a situation: instead of developing the same program from scratch for each different PLC brand, a unique program is stored using the PLCopen XML standard. When the program must be ported to a given PLC, an automatic tool, the postprocessor, generates the final code that has to be downloaded to the target PLC. In this paper the PLC tool adapter is presented and an example is shown for three of the major PLC brands present in the packaging industry such as Siemens, Schneider Electric and ABB.

Partial Inductance Model of Induction Machines for Fault Diagnosis

The development of advanced fault diagnostic systems for induction machines through the stator current requires accurate and fast models that can simulate the machine under faulty conditions, both in steady-state and in transient regime. These models are far more complex than the models used for healthy machines, because one of the effect of the faults is to change the winding configurations (broken bar faults, rotor asymmetries, and inter-turn short circuits) or the magnetic circuit (eccentricity and bearing faults). This produces a change of the self and mutual phase inductances, which induces in the stator currents the characteristic fault harmonics used to detect and to quantify the fault. The development of a machine model that can reflect these changes is a challenging task, which is addressed in this work with a novel approach, based on the concept of partial inductances. Instead of developing the machine model based on the phases’ coils, it is developed using the partial inductance of a single conductor, obtained through the magnetic vector potential, and combining the partial inductances of all the conductors with a fast Fourier transform for obtaining the phases’ inductances. The proposed method is validated using a commercial induction motor with forced broken bars.

Support vector machines optimization for steady state diagnosis methods of induction motors. A comparative study

Diagnosis of induction motors in industrial facilities is crucial to avoid unexpected outages that may lead a huge economic losses. Hence, this field has attracted a rising interest, in order to improve the accuracy of the fault diagnosis and to extend it to a growing number of different types of faults. Over the last decade, and recently, new diagnosis methods have being proposed in the technical literature which are solving some drawbacks of previous diagnostic methods and are improving the diagnostic procedure in terms of reliability, precision, and computational time, among others. In this paper, an optimization with support vector machine (SVM) of several diagnostic methods for accuracy improvement is presented, using a comparative study for finding the best and more optimized fault diagnosis system. A great amount of experimental tests have been developed to validate this comparative.