No abstract

1Qualitrol Hathaway Instruments, 2 Scottish Power Plc A method for harmonic measurements in EHV systems is proposed that is based on the measurement of currents in capacitor voltage transformers (cvt) and values of capacitor elements of cvts. This is the basis of a new technique, which uses current sensors installed within cvts. Simulation studies and field tests have confirmed the validity and practicability of the technique. In the new method the input terminal voltage is directly determined and no information about transfer function and burden is required. In order to verify the field test results, errors in the cvt output voltage recorded during tests were compensated by the use of the transfer function determined independently in transfer function tests for the same burden. A very good correlation was observed between the two results. The results of field measurements are also presented.

Following a power system fault, protection engineers have to analyse the protection scheme activity to ensure correct operation. To aid them in this task, data gathering systems are increasingly being fitted to power systems. However, during extreme operating conditions, the volume of data made available by these systems can be overwhelming. To help overcome this, knowledge-based systems have been developed and installed at ScottishPower's corporate headquarters to extract the relevant `information' from the supervisory control and data acquisition (SCADA) system `data'. The paper reports on the next stage of decision support for the protection engineers which builds upon the diagnosis provided by the knowledge-based systems. This enhanced support is provided by a model-based diagnostic system which provides automatic analysis of the available fault recorder data. This system utilises the most appropriate techniques employed by existing model-based diagnostic systems. A novel approach for handling tolerances associated with the operating times of protection scheme components is introduced, thus enabling modelling measurement inaccuracies to be dealt wit

Power system disturbance diagnosis is a complex and time consuming task, requiring skilled engineers. Existing intelligent systems which assist with data interpretation require the engineer to further interpret and collate generated information so that a comprehensive diagnosis can be produced. Furthermore, long-term extensibility and flexibility are also restricted. This paper presents a multi-agent approach to power system disturbance diagnosis where existing intelligent systems are wrapped up as intelligent agents and, through a process of inter-agent communication, collaborate to provide disturbance diagnosis. This novel approach enables timely automated disturbance diagnosis within a flexible and extensible architecture

No abstract