Naslov
Polynominal Identification of Strain Gauge Thermal Output for the Determination of Deformations with Neural Networks

Autori
Vujović, Igor ; Kuzmanić, Ivica ; Kulenović, Zlatan

Izvornik
Promet-Traffic-Traffico (0353-5320) 14 (2002), 6; 277-284

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, prethodno priopćenje, znanstveni

Ključne riječi
strain gauges; deformation; neural network; identification; thermal output

Sažetak
Strain gauges are used in different areas, especially in the design and development of new technical constructions and model testing. Also, strain gauges are incorporated in the functional part of many instruments and devices. They are mostly used as sensors in transducers design to measure such mechanical quantities as forces, moments, pressures, accelerations, etc. They have an imortant role in shipbuilding and marine transport in general. In this paper we have suggested and shown an approach to the identification of strain gauge thermal output curve on the example of a product available at the market. Neural network simulated in Matlab has been applied. The neural network has been adapted to simulate a real system with 10-9 order of magnitude error. As it is well known, strain gauges measure deformation of 10-6 order of magnitude. It is obvious that the network error cannot influence the measurement results because of its being smaller by three orders of magnitude.

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika

POVEZANOST RADA