An Ann Based Pi Controller for a Superconducting Generator

Abstract

The paper presents a coordinated artificial neural network (ANN) based proportional plus-integral (PI) controller for a superconducting generator (SCG). The ANN controller was trained using data groups, that covers the whole P-Q plane. These data groups obtained from a PI controller implemented on the governor control loop. To test the present control strategy, a fairly detailed nonlinear model of the SCG is used to assure accuracy and validity of the proposed control structure. The simulation results are presented in comparison with similar results which obtained using the conventional PI controller. The simulation results reveal that the ANN control design approach can be implemented to design controllers for the SCG and give results similar to those of conventional controllers.The natural growth of population renders continuous increase in electric power demand. One way of overcoming this problem is to develop SCGs. This machine offers a lot of advantages over the conventional synchronous machines and has the capability to supply greater base load with higher efficiency [1,2]. However, SCGs have a complex structure and require materials different fiom those normally used in conventional generators. This is due to that the high current density of the superconducting field winding obviates the need for any magnetic circuit in the SCG. For this reason all armature windings contemplated for SCG are of the air cored winding type. Also, the rotor incorporates in addition to the super-conducting field winding, a helium management system and two eddy-current screens [3,4]. The outer screen acts as a damper has time constant suitable for the damping purpose and the inner screen acts as an electromagnetic shield has a time constant sufficiently long to shield the field winding fiom time changing magnetic fields [5].

Artificial intelligence methods for applied superconductivity: material, design, manufacturing, testing, operation, and condition monitoring

“…Among these requirements is the increase of magnetic field in the excitation area between rotor and stator, the increase of the power density, reduction in size and weight, efficient cryogenic system with the minimized use of liquid cryogens, improved system reliability with minimized maintenance, a long system operation period of over ten years, and the availability of the operation of superconducting components over 99% of the time, which have higher priorities. AI techniques are powerful tools to deal with the challenges related to optimisation as well as condition monitoring [159][160][161][162][163].…”

Unsupervised Ann Based Pid Controller for a Superconducting Generator in a Multi-Machine Power System

“…Moreover, the magnitude and rate of change of the excitation current and field flux must not exceed certain limits, otherwise the superconducting element goes normal (quench) [12]. So, this renders the necessity of considering only the governor control loop to enhance the system performance [9]. Adding positive damping via the governor loop is very difficult and requires a great deal of attention [13].…”

“…Since 1989, ANNs methodology have capture and interest in electrical power engineering and it is promise to ensure effective ANN controllers, which can enhance system performance, if it properly trained. Many published researches have been developed ANN controllers for conventional generators and SCGs which trained using conventional schemes [9,10]. The object of this paper is to apply an alternative control approach using the ANN control strategy to improve the performance of the low inherent damping and high hunting frequency SCG in a multimachine power system.…”