| 1. |
- Balouji, Ebrahim, 1985, et al.
(författare)
-
Classification of Partial Discharges Originating from Multi-level PWM Using Machine Learning
- 2022
-
Ingår i: IEEE Transactions on Dielectrics and Electrical Insulation. - 1558-4135 .- 1070-9878. ; 29:1, s. 287-294
-
Tidskriftsartikel (refereegranskat)abstract
- A Partial discharge (PD) is a small, localized breakdown which often appears in high voltage insulation systems. PDs can be initiated within material defects such as voids or cracks when exposed to sufficiently high electric field levels. In this research work, we study and propose a machine learning approach to detect and classify PDs originating from multilevel pulse width modulation (PWM) waveforms as utilized in power electronic devices such as inverters and variable frequency drives. Due to the increased use of power conversion units, new types of higher magnitude PDs have been observed to increase insulation degradation. By creating a cavity placed at different locations within the test object and exposing it to PWM voltage waveforms with two different rise times, a total of 345660 PD events are recorded and organized into 10 different classes. The maximum PD amplitude, duration, the time distance between consecutive PDs, and the area under the PD are used as features for PD classification. A unique way of concatenating a sequence of the extracted features to capture the temporal dependence of consecutive PDs is also presented. It is observed that when creating a sequence of information from consecutive observed PDs, a significant increase in classification accuracy can be obtained. Trained classifiers based on ensemble bagged decision trees (DT) and long short-term memory (LSTM) architecture resulted in 95.3% and 98.5% average classification accuracy on test data respectively.
|
|
| 2. |
- Balouji, Ebrahim, 1985, et al.
(författare)
-
Partial discharge classification in power electronics applications using machine learning
- 2019
-
Ingår i: GlobalSIP 2019 - 7th IEEE Global Conference on Signal and Information Processing, Proceedings.
-
Konferensbidrag (refereegranskat)abstract
- A study of machine learning (ML) methods for classification of data from partial discharges (PDs) is described. A novel set of features are suggested and tested using an extensive set of machine learning based algorithms. The aim is to classify PDs occurring within insulation systems and power electronics devices (PED). Due to the increased use of pulse width modulation waveform (PWM) in PEDs, an increased insulation degradation has been observed due to a more intense PD exposure. This study aims to develop suitable tools to detect types of defects to facilitate diagnostics as well as to improve isolation system design. To evaluate the performance of ML based classification, several algorithms have been developed to detect and classify PDs from different kind of material defects with the aim to address the reason behind the appearance of partial discharges. Experiments with different PD source locations and volume of the defect and voltage rise time were investigated on an artificial cavity test object. Relevant signal features found important are for example the maximum magnitude, duration, the distance from polarity shift, the time distance between PDs and the absolute value of the area of the detected PD waveform. It has been observed that forming such PD features based on their time occurrence results in an accurate and generalized solution. With these features the best results were achieved with the deep learning LSTM architecture reaching a test accuracy of 98.3%. For industry applications, feature engineering is useful to reduce amount of data necessary to be analyzed by the neural network or ML algorithm.
|
|
| 3. |
- Bengtsson, Jörgen, et al.
(författare)
-
Method for measuring cylinder specific parameters in a combustion engine
- 2005
-
Patent (övrigt vetenskapligt/konstnärligt)abstract
- The invention concerns a method to measure parameters in the combustion chamber of a piston engine that comprises an ignition system. An alternating voltage is applied across the secondary winding of the ignition coil and the value of the current that arises in a measurement circuit that co-operates with the secondary winding is measured. The value of the current depends on the resistance (R1) of the measurement circuit, on the inductance (L1) and resistance of the secondary winding, and on the impedance of the combustion chamber, i.e. on its capacitance (C1) and its resistance. For example, top dead centre, the pressure in the cylinder, analysis of the ionic current and change of the burning time can be determined by means of the method
|
|
| 4. |
- Bengtsson, Tord, 1954, et al.
(författare)
-
Stochastic Detection of Partial Discharges
- 2013
-
Ingår i: IEEE Transactions on Dielectrics and Electrical Insulation. - 1558-4135 .- 1070-9878. ; 20:6, s. 2203-2211
-
Tidskriftsartikel (refereegranskat)abstract
- The traditional method to detect partial discharges relies heavily on the use of high-pass filters. Whereas this is a very effective approach for normal power voltage shapes, severe problems arise when partial discharges are to be detected under waveforms generated by modern power electronics. Therefore, a technique that utilizes the random occurrence of partial discharges for detection has been developed and used for semi-square voltages. This technique has evolved since the first publications and new applications have been found. The present paper introduces the technique and discusses the necessary analysis elements, requirements imposed as well as some examples intended to show possible application areas.
|
|
| 5. |
- Bergman, Becky, 1970, et al.
(författare)
-
Reflections on an Integrated Content and Language Project-Based Design of a Technical Communication Course for Electrical Engineering Students.
- 2013
-
Ingår i: Journal of Academic Writing. - 2225-8973. ; 3:1, s. 1-14
-
Tidskriftsartikel (refereegranskat)abstract
- Effective ways of teaching technical communication skills to engineering students have been much discussed. This article reflects on one setting, a first year course in Technical Communication at a university in Sweden, where electrical engineering teachers, language and communication teachers and student counsellors work in close, team-based cooperation using a project model which requires the students to analyse, implement and communicate technical problems. The paper discusses the change in this course - from an EAP course primarily prioritizing language training which ran parallel with a project course - to one unified ICL course. The progression is described through the changes in the organization of the course, and the article focuses on one learning activity: interdisciplinary tutorials on project reports. Through a pilot study where these sessions were video recorded and mapped, we conclude that the presence of different roles became an asset for the range of what the students see as relevant for their project report. In particular, the technical report genre was critically analysed, including problematic areas such as textual sequencing and display of technical problems; data visualisation and commentary; and referencing.
|
|
| 6. |
- Biswas, B., et al.
(författare)
-
Trends in Diagnostics and Monitoring of High-Voltage Insulation
- 2024
-
Ingår i: IEEE Electrical Insulation Magazine. - 0883-7554 .- 1558-4402. ; 40:4, s. 6-26
-
Tidskriftsartikel (refereegranskat)abstract
- This paper is a joint contribution from members of the IEEE DEIS Technical Committee for Diagnostics that highlights selected trends, challenges, and techniques in diagnostics and monitoring of high-voltage electrical insulation. Especially important are broad application areas including traditional generation, transmission and distribution, and renewable energy, along with emerging fields including new sectors in transportation electrification such as electromobility and electric aviation.
|
|
| 7. |
|
|
| 8. |
- Hammarström, Thomas, 1974
(författare)
-
A General Measurement Set-up Approach to Evaluate Insulation System Quality by Exploring PWM and DC Biased waveforms
- 2020
-
Ingår i: IEEE Transactions on Industry Applications. - 0093-9994 .- 1939-9367. ; 56:4, s. 3614-3624
-
Tidskriftsartikel (refereegranskat)abstract
- This article presents a novel test approach able to detect and evaluate PD characteristics when employing various wave shapes similar or close to what is used in different applications. The method covers the range from bipolar PWM with ultra-fast rise times to dc voltages with or without superimposed high -frequency content. Measurement results from tests on twisted pair aimed to resemble a random wound motor are presented. In particular, the influence of conductivity, permittivity, and rise time is explored by applying square- shaped waveforms. The PDs appearing after the unipolar flanks are compared with full -size bipolar voltage steps. One important observation presented is that when the zero voltage was reached and kept, the remaining charges within and surrounding the defect produced a PD pattern (PD echo) which resembles the charge amount present within the winding. Additionally, PWM shaped voltage waveforms superimposed on dc voltages present advantages when further evaluating the influence of material properties. Different step sizes are employed to represent dc voltage with or without superimposed voltage ripple to elucidate the influence on PD characteristics also on HVdc like applications. It can be concluded that higher frequency harmonics will result in considerably larger PD exposure. This article shows that during material evaluation, important common properties and guidelines can be found between these different waveform extremes, suggesting that a more efficient insulation coordination can be obtained selecting the most appropriate or time- efficient evaluation method.
|
|
| 9. |
- Hammarström, Thomas, 1974
(författare)
-
A Measurement Set-Up Method to Evaluate Motor Winding Quality and Winding Insulation by Exploring Discharge Characteristics
- 2018
-
Ingår i: 2018 21st International Conference on Electrical Machines and Systems (ICEMS). - 9788986510201 ; , s. 678-683
-
Konferensbidrag (refereegranskat)abstract
- The development of new semiconductor materials such as SiC and GaN based designs introduces shorter rise times as well as increased carrier frequencies in power electronic applications. One very common reason for motor winding failures is due to the presence of partial discharges (PDs). This paper presents a test method to detect and evaluate the presence of PDs employing wave shapes similar or close to what is used in applications. The waveform is generated by applying a novel three-level inverter circuit. The set-up and detection method can for example be used to evaluate the influence of rise times, reflections and insulation conductivity on simple twisted pair test samples or complete windings. Initial measurement results from tests on twisted pair aimed to resemble a random wound motor are presented. In particular, the influence of conductivity, permittivity and rise time is explored by applying square shaped waveforms. The PDs appearing after the unipolar flanks are compared with full size bipolar voltage steps. When the zero voltage was reached the remining charges within and surrounding the defect produced a PD pattern (PD echo) resembling the charge amount within the winding. Tests showed a significant difference due to the choice of material. The introduced measurement approach maintains the advantages of using impulses to evaluate windings but complemented with charge decay measurement.
|
|
| 10. |
- Hammarström, Thomas, 1974, et al.
(författare)
-
CHANGES IN PD PROPERTIES WITH DECREASING VOLTAGE RISE
- 2011
-
Ingår i: The Proceedings of 17th International Symposium on High Voltage Engineering. - 9783800733644 ; , s. D-013
-
Konferensbidrag (refereegranskat)abstract
- The increased use of power electronic components in power systems requiresunderstanding of how rapidly rising voltages affect insulation systems. One importantaspect of this challenge is to measure partial discharges, PDs, which are considered asbeing a sign of insulation weakness and can affect its life considerably. This paperpresents a continuation of our earlier investigations on the difference in PD behaviourwhen voltages characterized by different rise times are applied. The results indicate thatthe effect on the insulation system is indeed dependent on the voltage wave shape.Applying square-like voltages to a cavity with dielectric insulated electrodes significantlyaffects the discharge amplitude, its rise time, the inception voltage and the distributionshape. It is further examined how the dimensions of the cavity affect the PDcharacteristics and the investigation shows that PD amplitude increases while its durationdecreases for shorter voltage rise times, being indications of a possible change in thedischarge mechanism. This in turn can lead to faster deterioration and reduction ofinsulation service life. To illustrate the degradation process, microscopic images showhow shorter rise times affect differently the cavity surface deterioration, which isconsistent with the other observations.
|
|
