A DISO NEURO-FUZZY CONTROL FOR DC SERVODRIVES

Journal Title: ANALELE UNIVERSITATII DIN CRAIOVA - Seria Inginerie electrica - Year 2011, Vol 35, Issue 35

Abstract

The paper follows a previous work that made the synthesis of a neuro-fuzzy controller with double input and single output for positioning systems with DC motors. Now, the new system is put into operation using models and simulations for proving its qualities. Briefly, several test will certify that this single neuro-fuzzy controller (for both the position and the speed) is able to ensure good or very good performance in steady-state and dynamic regimes for a wide variety of conditions concerning: the motor parameters, the load characteristics (torque, disturbances, inertia), the reference cycle (structure, values, duration) and the sampling time. The global results for both the control structures (standard – that generated the training data and fuzzy) are compared. The next stage is to prepare the tools for an on-line implementation. Several formulas will be suggested for the real-time control using 2D and 3D look-up-tables by direct search and by interpolation. The author sustains an on-line timing based on imbricated hardware / software interrupts, that will be implemented using a low-cost microcontroller. The experimental platform is presented as well as some real-time results for the main variables of the system and by recording the on-line timing with a logic analyzer. Most of the ideas, formulas and implementation procedures presented here are useful for other systems and applications.

Authors and Affiliations

Dan MIHAI

Keywords

Related Articles

Simulation of Regenerative Braking at an Electrical Scooter

The paper deals with MATLAB simulations of the regenerative braking for an electrical scooter. Various modern possibilities to achieve regenerative braking at light electric vehicles are approached firstly. Afterward...

Contactless Power Transfer – Theoretical Principles and Fields of Applications

Contactless energy transfer 1 has indubitable advantages in modern technique given by the lack of electro-mechanical contacts predisposed to pitting and failure: use in aggressive or explosive (no sparks) environments, a...

The Study of Asynchronous Motors Behavior in Artificial Load

The performances and heating of asynchronous motor are established to nominal load into a testing laboratory. Only in this way can be created real life working conditions, but there are also two big disadvantages: the ne...

Optimal Shape of DC Electromagnet

In this paper is proposed an optimal geometrical shape for a DC electromagnet by combining the design of experiments (DOE) and 2-D FEM. The optimization prob-lem takes into account four geometrical parameters (the coil s...

Design and Performance of the Voltage Control Loop in Induction Heating Systems with L-LC Resonant Inverters

In the induction heating systems1 with voltage source inverter and resonant L-LC load, different ways to control the power transferred to the heated piece are possible. This paper takes into consideration separate contro...

Download PDF file
  • EP ID EP86241
  • DOI -
  • Views 203
  • Downloads 0

How To Cite

Dan MIHAI (2011). A DISO NEURO-FUZZY CONTROL FOR DC SERVODRIVES. ANALELE UNIVERSITATII DIN CRAIOVA - Seria Inginerie electrica, 35(35), 146-151. https://europub.co.uk/articles/-A-86241