A Fuzzy Logic Protection Scheme Implemented in a DC Traction Substation Grounded Track zyxwvuts E. Beaud Rectifier Transformer 3 zyxwvu - 4kV DC Overhead Line 2o zyxw km Department of Electrical & Electronic Engineering University of Stellenbosch, Private Bag X1, Matieland, 7602, South Africa zyxwvutsr Abstract - This paper proposes the combination of a fuzzy controller with a conventional system protection controller to give the system a measure of ride-through capability. The fuzzy controller forms a layer of protection between the normal operating range and the conventional protection scheme offering greater system robustness. 1. INTRODUCTION Fuzzy Control has, since its inception in the sixties [l], been used in various applications in industry. It has proven to be very useful in the control of non-linear and time varying systems[2,3]. Fuzzy logic has by its name, given the world the impression that there is a measure of uncertainty in its response. Its widespread use in industry has proven its reliability and versatility. The industry has been reluctant in applying fuzzy logic for the protection of equipment and has rather opted for large safety margins. This leads to a utilization of equipment that is less than optimal. All systems used in industry applications have an area of safe operation. This might vary in complexity from a simple motor drive that is limited to a certain maximum speed, to complex systems that have numerous currents, voltages, speeds, etc. that cannot exceed specified limits for each parameter. In a typical system a protection scheme will prevent the system from exceeding its specified limits. Normally the protection strategy is divided into layers of protection as shown in Fig. 1. ----------------- Sub-system Protection _---__----___ Component Protection Fig 1 Layers of protection in a standard system The protection would operate from protecting individual components, then individual sub-systems through to the protection of the system as a whole as well as the operation of the system. At every level there are parameters like speed, Voltage and Amperes that may not be exceeded An error in one of these parameters would cause a shutdown of the appropriate subsystem or the total system according to the error detected. 0-7803-4340-9/98/$10.00 zyxwvuts 0 1998 IEEE. 1120 J.H.R. Enslin Department of Electrical & Electronic Engineering University of Stellenbosch, Private Bag X1, Matieland, , 7602, South Africa Some errors , like peak currents, that could cause damage over a long period of time would be harmless for short durations [4]. If the weight of the error could be used in the determination of the reaction, the protection scheme could be optimized to allow the system to operate at its maximum safe operating conditions. In the application introduced in this paper, a fuzzy controller would meet the requirements as it is not inherently an ordoff controller, but rather tries to emulate the human mind by evaluating the input conditions and reacting accordingly. It would therefore weigh input conditions and act according to the individual weights of the errors to protect the system. This paper proposes the application of a fuzzy logic protection strategy, as an example, to a dc traction substation and train simulator prototype system. 2. SYSTEM DESCRIPTION The traction system of South Africa consists mainly of 3.3 kV overhead lines feeding dc traction locomotives. At places there have been installed 22 kV ac lines, but the majority is still 3.3 kV dc. The dc traction system consists of substations spaced along the dc overhead lines at intervals of about 20 km (Fig. 2). Traction I I I Substation AC-SUDD~V: Rectifier Transformer I I ' I Fig. 2 : DC Traction system in South African Railways