Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
8 RF Emission-Based Sensors RF emissions from the power network relate the condition of the network to the frequency content of those emissions. Figure 6 provides the block diagram of the RF emission-based sensor. The sensor has been tested at the Greater Cleveland Regional Transit Authority (GCRTA) DC railway system. In Cleveland, Ohio, the interurban/light rail lines use overhead wires. Similarly, the heavy rail line also uses overhead wires, instead of a third rail. This is due to a city ordinance intended to limit air pollution from the large number of steam trains that passed through Cleveland between the east coast and Chicago. Trains switched from steam to overhead catenary electric locomotives at the Collinwood Rail Yards, about 10 miles (16 km) east of Downtown Cleveland, and at Linndale on the west side. When Cleveland constructed its rapid transit (heavy rail) line between the airport, Downtown Cleveland, and beyond, it employed overhead catenary technologies similar to those used by the railroads, and used the railroad electrification equipment that remained after the railroads switched from steam to diesel locomotives. Light and heavy rail public transit systems share about 3 miles (4.8 km) of track along the Cleveland Hopkins International Airport Red line (heavy rail), the Blue and Green interurban/light rail lines between Clevelandâs Union Terminal, and just past the East 55th Street station, where the heavy and light rail line tracks separate. 3.1 Data Structure and Interface for the Sensor Data The Exacter Real-Time Failure Signature Sensor evaluates RF emissions from deteriorated electrical equipment. The sensor is installed within the train. Data is automatically uploaded from the sensor via cellular data networks to cloud-based servers and includes magnitude and Global Positioning System (GPS) location data of the sensed failure signature, the Event. The Event information is used in the geospatial analysis of the location of the failure signature emission source by trade secret algorithms that reside in the Exacter Cloud-based servers. 3.2 Data Event A data Event is recorded by the Exacter Sensor whenever RF emissions correlate to the Exacter Failure Signature Library. The RF emissions were collected by the sensors during spring and summer 2016. Figure 7 shows the Events recorded. The data collected represents over 67,000 emission samples taken of background RF emissions. The data includes three sets of data analyzed using machine-learning algorithms. The data sets are DC arcing that is not related to deteriorated equipment. This occurs when- ever the pantograph contacts connectors on the trolley wire or crosses a trolley wire separation; AC arcing leakage and tracking on AC electrical systems adjacent to the railway or that the train crosses under during its route; and DC arcing leakage and tracking, which is the target C H A P T E R 3
RF Emission-Based Sensors 9 data that indicates the deterioration of the DC electrical system and opportunities for predictive maintenance to avoid electrical outages. 3.3 Data Security Data is uploaded to the Exacter Cloud for geospatial analysis and is encrypted using the North American Electric Reliability Corporation Critical Infrastructure Protection plan (NERC CIP) Version 6 compliant, 128-bit encryption algorithms designed for electric infrastructure cyber security. Figure 6. RF emission-based sensor components. Figure 7. Data events.
10 Guidebook for Detecting and Mitigating Low-Level DC Leakage and Fault Currents in Transit Systems 3.4 Sensor Location The sensors are in the attendant compartment of the lead train engine. Figure 8 shows the prototype installation of two sensors and associated DC to DC isolation and protection equipment. One sensor is used to detect DC RF emissions. The second sensor is used to detect AC RF emissions. Including both AC and DC sensors in the initial data collection allowed the machine-learning algorithms to eliminate any contamination of the DC emission data by identifying and eliminating the AC emission data. Future installations will not require the inclusion of the AC emission sensor. Once the installation of the sensors is complete, there is no interaction required from the train operator. The sensors operate autonomously and activate during train movement. Figure 9 shows the RF emission, train motion, and GPS antenna systems mounted on the roof of the train. The entire system is powered from the train DC voltage supply and is isolated using DC to DC converters for isolation and protection of the train DC supply system. A specification sheet for the prototype sensor is included as Figure 10. 3.5 Failure Signatures Key to the success of the project, and the focus of the initial phase of the project, is to identify and record Failure Signatures (FSs). The FSs are added to the Exacter Failure Signature Libraryâ¢ and are the triggering emission to establish a FS Event. The computer dashboard in Figure 11 is the evaluation tool that Exacter data scientists use to evaluate the many RF emissions detected during a train trip and to establish the specific signature to be used to trigger the sensor for an RF emission source location study in the final sensor design. Shown in this figure is the evalu- ation of RF emission correlation to a design FS from many non-correlating Events recorded in the initial emission studies. Figure 8. Installation.
RF Emission-Based Sensors 11 Figure 9. Antenna systems. The characterization of DC arc emissions requires that the sensor process data in real time while the train is moving. The DC arc emission has three characteristics to create a FS. The cor- relation of these three characteristics, shown in Figure 12, allows for precisely evaluating a DC arcing, leakage, or tracking Event, and distinguishes the analyzed emission from DC sparking Events and AC system Events. Of the 67,000 emission studies that were run, 22 specific FS candidates were established. These were then analyzed by Exacter Cloud geospatial algorithms, and the emission source locations were established. The map in Figure 13 shows these initial locations, indicated by the black and red Event markers. When Events are geospatially located, they are renamed Maintenance Groups. Maintenance Groups are indicators of locations that should be visited by maintenance personnel to evaluate the electrical equipment at the location for maintenance or replacement. 3.6 Group Maintenance Reports Group Maintenance reports are automatically generated as maintenance locations are resolved. The reports are provided in a variety of formats, including Shapefile (SHP) files for direct import to standardized geographic information systems (GISs), comma-separated values (CSV) files, and Keyhole Markup Language (KMZ) files. A sample KMZ file is shown in Figure 14 using the free Google Earth program (a program copyrighted by Google Corporation). The latitude and longitude of the FS emission source are visible. In the initial assessment of the 22 Maintenance Groups located during the initial phase of the project, the maintenance issue was not clearly identifiable in the field. The research team learned that the ultrasonic equipment used in locating AC system maintenance issues, such as cracked or contaminated insulators, was not effective when applied to a DC electrical system. A localized DC emission detector has been designed and will be used for field trials in the next phase of the project.
12 Guidebook for Detecting and Mitigating Low-Level DC Leakage and Fault Currents in Transit Systems Figure 10. RF emission-based sensor specifications.
RF Emission-Based Sensors 13 Figure 11. Computer dashboard. Figure 12. Failure Signature (FS) correlation.
14 Guidebook for Detecting and Mitigating Low-Level DC Leakage and Fault Currents in Transit Systems Figure 13. Failure Signature black and red event markers. Figure 14. Sample KMZ file.