S.U. Uvaysov1, V.V. Chernoverskaya2, Nguyen Van Tuan3, Nguyen Viet Dang4

1-4 RTU MIREA (Moscow, Russia)

The article considers an original method for diagnosing analog circuits of radio electronic devices, which is a combination of frequency analysis based on the Seshu-Waxman approach with the Kohonen algorithm for further recognition of hidden defects. The frequency response is examined at circuit test points, and the transfer function is used to determine test diagnostic frequencies.

The purpose of the study is to develop a method for technical diagnostics of electrical circuits of radio electronic components, which increases the reliability of diagnosis by combining traditional control methods with machine learning technologies and the use of artificial neural networks.

Main results of the study: developed method for technical diagnostics of electrical circuits of radio electronic components, which increases the reliability of diagnostics by combining traditional control methods with machine learning technologies and the use of artificial neural networks.

The structure of the proposed method has been developed with a description of the functionality of logically connected nodes and a sequence of research stages. A schematic diagram of a radio electronic device, created in the NI Multisim environment, and a family of LAFC for various technical states of the REU are presented. During the computational experiment, a set of test frequencies and values of the parameters of the circuit elements was obtained. On the basis of the obtained values, a database of REU malfunctions was formed, which is a two-dimensional array and serves for subsequent training of an artificial neural network (ANN). The ANN training procedure based on the Kohonen algorithm is implemented in the Python programming language. The visualization of the process of classifying defects by an artificial neural network based on test vectors is given. Circuit design, simulation and engineering analysis were carried out using the specialized NI Multisim environment, the development of the ANN was carried out using the high-level Python programming language, connecting additional libraries and tools.

Practical significance of the study: the materials obtained as a result of research are introduced into the methodology for designing radio-electronic devices.

Uvaysov S.U., Chernoverskaya V.V., Nguyen Van Tuan, Nguyen Viet Dang. Application of the Seshu-Waxman approach and the Kohonen algorithm to the problem of electrical diagnostics of analogue circuits. Radiotekhnika. 2022. V. 86. № 12. P. 79−89.
DOI: https://doi.org/10.18127/j00338486-202212-07 (In Russian)

1. Andrejchikov A.V., Andrejchikova O.N. Intellektual'nye informacionnye sistemy i metody iskusstvennogo intellekta: uchebnik. M.: INFRA-M. 2021. 530 s.
2. Valjaev E.A. Nejrosetevoj kaskad na osnove samoorganizujushhihsja kart Kohonena. V kn.: Sbornik izbrannyh statej po materialam nauchnyh konferencij GNII «NACRAZVITIE». SPb: GNII «Nacrazvitie», 2019. Ch. 2. S. 192−204 (In Russian).
3. Volovikova E.V. Metod diagnostirovanija radiojelektronnyh funkcional'nyh uzlov po jelektricheskim harakteristikam s uchetom temperatur komplektujushhih jelementov: Avtoref. diss. … kand. tehn. nauk. M.: MIJeM. 2010. 152 s
4. Besshejnov A.V., Idzhelliden S.B. Instrumental'noe sredstvo dlja formirovanija bazy harakternyh neispravnostej radiojelektronnyh ustrojstv. Sb. nauch. trudov nauch.-tehnich. konf. studentov, aspirantov i molodyh specialistov instituta, posvjashh. 40-letiju MIJeM. M.: MIJeM. 2002 (In Russian).
5. Gorodov V.A. Jelektricheskij kontrol' pechatnyh plat i uzlov. Jelektronika: Nauka. Tehnologija. Biznes. 2004. № 7. S. 68–71 (In Russian).
6. Konoval'chuk A.S. Kompleksnoe modelirovanie jelektricheskih i teplovyh processov v analogovyh mikrojelektronnyh uzlah: Avtoref. diss. … kand. tehn. nauk. M.: MIJeM. 1987. 230 s (In Russian).
7. Bjendler Dzh.U., Salama A.Je. Diagnostika neispravnostej v analogovyh cepjah. TIIJeR: Per. s angl. 1985. T. 73. № 8. S. 3587 (In Russian).
8. Uvajsov S.U., Chernoverskaja V.V., Dao An' Kuan, Nguen Van Tuan. Algoritm Kohonena v zadachah klassifikacii konstruktivnyh defektov pechatnyh uzlov. Rossijskij tehnologicheskij zhurnal. 2021. T. 9. № 4. S. 98−112. https://doi.org/10.32362/2500-316X-2021-9-4-98-112 (In Russian).
9. Gorbachenko V.I., Ahmetov B.S., Kuznecova O.Ju. Intellektual'nye sistemy: nechetkie sistemy i seti: uchebnoe posobie dlja vuzov. M.: Jurajt, 2019. 105 s.
10. Hajkin S. Nejronnye seti: polnyj kurs. M.: Vil'jams, 2006. 1104 s
11. Andrejchikov A.V., Andrejchikova O.N. Intellektual'nye informacionnye sistemy i metody iskusstvennogo intellekta: Uchebnik. M.: INFRA-M, 2021. 530 s.
12. Valjaev E.A. Nejrosetevoj kaskad na osnove samoorganizujushhihsja kart Kohonena. V kn.: Sbornik izbrannyh statej po materialam nauch. konf. GNII «NACRAZVITIE». SPb: GNII «Nacrazvitie». 2019. Ch. 2. S. 192−204
13. Kohonen T. Samoorganizujushhiesja karty: Per. s angl. M.: BINOM. Laboratorija znanij. 2008. 655 s. 
14. Zhang C., He Y., Yuan L. Analog circuit incipient fault diagnosis method using DBN based features extraction. IEEE Access. Apr. 2018. V. 6. Р. 23053-23064.
15. Stovbun I.Ja., Uvajsova S.S. Primenenie podhoda Seshu-Uoksmena dlja formirovanija jeffektivnyh testovyh vozdejstvij. Sb. materialov konf. «Innovacionnye, informacionnye i kommunikacionnye tehnologii». M. 2016. T. 1. S.436-438 (In Russian).