Complex Event Processing in Power Distribution Systems: A Case Study Debnath Mukherjee Deepti Shakya Prateep Misra Tata Consultancy Services Limited Plot A2, M2 & N2, Sector V, Block GP Salt Lake Electronic Complex, Kolkata, India 700091 {debnath.mukherjee, deepti.shakya, prateep.misra}@tcs.com Abstract Complex Event Processing (CEP) is an emerging discipline. This paper focuses on the application of CEP for fault detection and classification in 11kV radial distribution system using data collected from a Phasor Measurement Unit (PMU). The analysis has been done by monitoring the electrical quantities in the 11kV radial distribution system simulated using Matlab Simulink. The PMU is placed at the substation and transmits data to the Command and Control Center. The data is analyzed to identify the signatures of different types of faults and based on that rules have been designed to categorize them. An architecture stack has been designed based on a commercial CEP product (Tibco Business Events) to implement the fault detection. In this paper we present the architecture, the 11kV distribution system simulation and share our experience about fault detection. The paper outlines the categorization and analysis of various types of faults like Single Line to Ground (SLG) fault, Double Line to Ground (DLG) fault, and Three phase (3) fault using CEP software. Thus it shows a real life application of CEP software for fault classification with low computational time. Keywords Complex Event Processing, Phasor Measurement Units, Architecture, Simulation, fault detection, fault classification 1 Introduction Complex Event Processing (CEP) is an emerging technology which can be used to detect interesting patterns amongst events being received in real time. There are two related fields: Event Stream Processing is concerned with time-ordered sequence of events. CEP is of wider scope which includes partially ordered set of events known as “event clouds” [1]. In this paper we describe how CEP can be used for detection and classification of faults in 11kV distribution system using PMU. Fault detection and clearing is important for any network from safety and reliability perspective. In particular, we describe techniques that can be applied to distribution systems that monitor phasors using Phasor Measurement Units (PMU). The unique contribution of this work is efficient fault detection and categorization using CEP technology. 2 Related Work Current fault detection technology is based on current transformer based relaying. The relay co-ordination ensures that the impacted zone is minimum. The shortcoming of this is relay setting and relaying accuracy may differ, sometimes not providing optimum fault clearance. The authors are not aware of any reported fault categorization and detection work based on CEP technology. CEP platforms for phasor data concentrators and stream processing have been reported in [2]. However it does not mention specific use of CEP for fault detection and classification. Analysis of cause of faults and fault location is being considered in future work. 3 Phasor Measurement Units PMUs are widely used in transmission networks for applications on Wide area monitoring and protection, Voltage instability analysis and prediction. Suitably engineered PMUs can also be integrated in to distribution level IEDs (Intelligent Electronic Devices). The paper examines such distribution level PMUs for fault detection and classification. 4 Problem Statement The problem under consideration is detection of occurrence and type of faults in the 11kV radial distribution system. This distribution system is simulated using Matlab Simulink. Three phase, 13 bus balanced radial distribution system is simulated in which a PMU is placed at bus 1. The PMU is simulated using the three phase measurement block of Simulink library. Appendix A shows the system details. Table 4 mentioned in appendix A shows the line and load data for the test system. The system is simulated at frequency of 60Hz. In this simulation, different types of faults are created at different locations. At bus 1, three phase voltage and current waveforms are sampled with sampling frequency of 1kHz, in form of magnitude and phase angles. The PMU sends this data to Command and Control station where it is processed for fault detection and classification using CEP software. The format of the transmitted synchronized data follows the standard mentioned in IEEE Standard C37.118 [3]. PMU aggregates the signals i.e three phase voltage and current in terms of magnitude and phase angles. These data are then sent to a Command and Control station where the data is processed for fault detection and classification using CEP software. The format of the transmitted synchronized data is similar as mentioned in [3]. The captured data is then used to detect the type of fault occurred in the distribution system. 5 Using PMUs to detect faults Let A, B, C and G represent phase a, b, c and ground. Different types of fault that can occur in the distribution system are: Single Line to Ground Fault (SLG) i.e. AG, BG and CG. Line to Line fault (L-L) i.e. AB, BC and CA. Double Line to Ground Fault (DLG) i.e. ABG, BCG and CAG. Three phase Fault (3L) i.e. ABC.