Bulanık mantık yöntemi kullanılarak yakın alan probunun optimizasyonu

Abstract

Elektromanyetik Uyumluluk(EMU),(EMC)-Elektromanyetik Girişim(EMI)-Bağışıklık- Yakın alan probu-Bulanık Mantık, Bulanık çıkarım, SC (Soft Computing) sensor Elektromanyetik yakın alan ölçümünde kullanılan probların endüktanslarının hesaplanmasında ve prob parametrelerinin tasarımında genellikle ampirik formüller kullanılmaktadır. Bu formüller prob yarıçapının veya prob kenar uzunluğunun iletken yarıçapına oranının yüksek olduğu değerlerde daha doğru sonuçlar vermekte ancak bu oranın küçük olduğu değerlerde ise hatanın yüksek olduğu görülmektedir. Bu çalışmada, prob yarıçapının iletken yarıçapına oranına bakılmaksızın geçerli olabilecek ve hata oram daha düşük olan bir formül geliştirilmiştir. Tasannu yapılan 56 adet prob üzerinde ölçümler yapılmış ve Bulanık Mantık Yöntemi kullanılarak yeni, düzeltilmiş formüller, Halka prob için; Lhalka = N2 x 0. 1 62 x /(2.3 x log - - 2A5)(jıH), a Kare biçimli prob için; ZsJV2x0.162x/(2.3xlog- -2.85)(/tff) a şeklinde elde edilmiştir. Daha önce kullanılan ampirik formüller ile bulanık mantık yöntemiyle geliştirdiğimiz formüllerin endüktans hesaplamaları ve ölçümleri karşılaştırılmıştır. Geliştirilen formüllerin, özellikle uygulamada kullanılan yakın alan problarının endüktanslarının 20 nH-70 uH gibi büyük bir aralıkta hesaplanmasında %1'den küçük bağıl hata verdiği görülmüştür.

OPTIMIZATION OF NEAR FIELD PROBES BY USING FUZZY LOGIC Key Words: EMI/EMC, EM Immunity, Near Field, Standard Probe, Fuzzy Logic Inference (FIS), Resistive Loading, Soft Computing Sensors Electromagnetic Compatibility, or EMC, is essentially the opposite of EMI.- EMC means that the device is compatible with (i.e., no interference caused by) its EM environment, and it does not emit levels of EM energy that cause EMI in other devices in the vicinity. Near Field Electromagnetic measurements are required for EMI/EMC studies. They are also required for susceptibility analysis of prototype of electronic devices. At the beginning of design state of device, electromagnetic tests must be provided to eliminate fatal problems at producing state. Near field studies must be carried out for shielding printed circuit boards and other circuits entirely from outer space. It can be achieved by utilizing screen properly. These studies must be always made in the near field. According to EM principles, interference occurs as radiated and conducted. Basically return current loops(paths) or some other paths radiate the unwanted RF energy. The best way to sense this energy has been known using a proper loop antenna. The loop probe can be used both as a transmitter and receiver. If it is loaded resistively, desired flat response between low and high cut off frequencies can be obtained. To get accurate results, its physical and electrical parameters must be chosen properly. The most important electrical parameter is inductance of such loop shaped conductor. Namely, the inductance must be calculated accurately for the best results. So, one can calculate the inductance without making a complex measurements. Common inductance equations are in general form and they are derived for general calculations. However loop probes have narrow physical range; they have a few- centimeter diameters, a few-turn, and with thin conductors. So, a certain equation can be derived with more accuracy. It can be managed by using conventional empirical derivations or by using modern optimization techniques. At present, Fuzzy Logic Algorithm is often used in every engineering discipline. Since every physical event has got stochastic uncertainty. FL works with linguistic variables such as "little-mid-much-etc." It is a soft computing(SC). SC is an emerging approach to computing. Its characteristics are "new application domains- fault tolerances-real world applications". So we use loop probes as soft sensors. Because of this we have to follow soft computing methodology. Fuzzy algorithm is in SC. In this way, stochastic uncertainty can be eliminated. Required rule base was provided from experimental results. For this, 56 different probes are made in very xivnarrow physical range. They were used to create rule base and data base. The membership function for FL can be created by using these bases. Then experimental results and Fuzzy Inference System(FIS) outputs are compared, some useful tables and graphs are plotted. Finally, error surfaces are plotted to see how our estimation is is close to the observed data. And a new empirical equation is developed to calculate the inductance of such probes. For validation of test results, calibrated commercial EM instruments were used. The graphs were generated by using some software. The main goal of this study was derivation of a new equation. At the end of the study, two new equations symbolized FISs were created. The one of them is for circular loop probes, and the other is for square loop probes.