Soft Fault Detection in Analog Electronic Circuit Using Three Types of Artificial Neural Network

Analog electronic circuit condition monitoring is gaining importance in electronic industry because of the need to increase product quality and to decrease the possibility of production failure due to faulty electronic component and circuit. Although often the visual inspection of the measured circuit response is adequate to identify the faults, there is a need for a reliable, fast, and automated procedure of diagnostics

The automated diagnosis and detection of faulty electronic circuit can be conducted by an Artificial Neural Network (ANN) [2-8]. The applications of ANN are mainly in the areas of Intelligent Agents, Monitoring and Warning Systems, Process Automation and Intelligent Personal Assistant because of their high accuracy and good generalization capability. Multilayer perceptrons (MLP) and radial basis functions (RBF) are the most commonly used ANN though interest in probabilistic neural networks (PNN) is also increasing recently [1]. The main difference among these methods lies in the ways of partitioning the data into different classes.
Though in the area of analog electronic circuit condition monitoring RBF was experimented [2]. It is important that the others type be experimented for comparing the performance of three types of the ANN and for finding the best performance of soft fault detection in analog electronic circuit.

The development of a diagnostic system using ANN consists of the following main steps: definition of faults of interest, selection of an optimal set of measurements, feature extraction (generation of circuit under test signatures), the choice of type and topology of the ANN, training and testing the ANN. In the diagnostic process of analog electronic circuit, the measured circuit response is compared with the signatures corresponding to each potential fault condition stored in the dictionary. By a fault in analog electronic circuit we mean any change in the element value which can cause a failure of the circuit.

In this research, collecting the references of diagnostic system using ANN is the first step. By collecting the references we can compare the results of research that have been done. The next step is designing and implementing the diagnostic system using three types of the ANN. Testing the system for each type of the ANN is very important for comparing the system performance. Performance comparison of the system is indicated by percentage of test success for each type of the ANN.

The aim of this research is to develop the ANN and dictionary approach for soft faults location in electronic circuits at the component level using three types of the ANN and to compare the performance of the three types of artificial neural network (ANN). The best performance of soft fault detection in analog electronic circuit using three types of the ANN will be found by evaluating experiment results. The result from this research can be proposed and implemented in the electronic industry for increasing product quality and decreasing the possibility of production failure due to faulty electronic component and circuit.

Bibliography:
[1] Samanta, B., al-Balushi, Khamis R., al-Araimi, Saeed A., Bearing Fault Detection Using Artificial Neural Networks and Genetic Algorithm, EURASIP Journal on Applied Signal Processing 2004:3, pp. 366–377.
[2] Wojciech Toczek, Michal Kowalewski, A Neural Network Based System for Soft Fault Diagnosis in Electronic Circuits, Gdansk University of Technology Faculty of Electronics Telecommunication and Informatics, 2005.
[3] Materka A., Strzelecki M.: Parametric Testing of Mixed-Signal Circuits by ANN Processing of Transient Responses, Journal of Electronic Testing: Theory and Applications, 9, 1996.
[4]Tang, H.; Mack, A dictionary approach to fault location in linear circuits, R.J.Circuits and Systems, 1991., IEEE International Symposium on Volume , Issue , 11-14 Jun 1991 Page(s):2064 – 2067 vol.4.
[5] Bernieri A., Betta G., Liguori C.: On-line Fault Detection and Diagnosis Obtained by Implementing Neural Algorithms on a Digital Signal Processor, IEEE Transaction on Instrumentation and Measurement, vol. 45, no. 5, 1996.
[6] Catelani M., Fort A.: Soft Fault Detection and Isolation in Analog Circuits: Some Results and a Comparison Between a Fuzzy Approach and Radial Basis Function Networks, IEEE Transactions on Instrumentation and Measurement, vol.51, no. 2, 2002,.
[7] Stopjakova V., Malosek P., Micusik D., Matej M., Margala M.: Classification of Defective Analog Integrated Circuits Using Artificial Neural Networks, Journal of Electronic Testing: Theory and Applications, vol. 20, 2004.
[8] Litovski, V., Andrejevic, M. and Zwolinski, M. (2006) Analogue electronic circuit diagnosis based on ANNs. Microelectronic Reliability, 46 (8). pp. 1382-1391. ISSN 0026-2714.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: