Proc. of STECH 2015 © 2015 The Japan Society of Mechanical Engineers  The International Symposium on Speed-up and Sustainable Technology for Railway and Maglev Systems November 10-12 2015, Chiba, JAPAN Identification of a ‘Driver Culture’ and its effect on energy consumption on a DC rail network R. Ellis*, P. Weston*, S. Hillmansen*, I. Jones** * The University of Birmingham, Edgbaston, B15 2TT, United Kingdom E-mail: rxe297@bham.ac.uk ** Head of Engineering, Merseyrail United Kingdom Abstract Performance analytics have successfully been implemented in sports, education and business & commerce. This paper presents data analysis that is intended to reduce energy consumption in the rail sector by evaluating the performance of drivers on a 750v DC network in Great Britain. Services on the Merseyrail Northern Line, to and from Hunts Cross to Southport, were picked out from data recorded on an instrumented British Rail Class 508, for the three month period September to November 2011. Aggression metrics during the acceleration period and deceleration period were classified for each run as well as energy consumption and running time, this was then used to group runs into distinct clusters. Clustering represents driver performance for the three month period which can be used to demonstrate a driver culture that exists within a relatively small group of driver personal. Driver instruction can be improved by evaluating performance data which can be used to reduce energy consumption by 10% on the observed line. Key words : Energy reduction, Data/Performance analytics, Big data 1. Introduction Railways offer the most efficient form of land based transport owing to the low rolling resistance of the wheel to rail contact (Hillmansen & Roberts, 2007). However a rail network is likely the largest consumer of energy within the area in which it operates (Acikbas & Soylemez, 2007). Energy reductions can be made by improving the design of the traction system and increasing the aerodynamic shape of the vehicle. Infrastructure design such as track alignment, gradient and banked curves also significantly contribute to reduction of energy consumption (Liu & Golovitcher, 2003). However, choice of rolling stock and infrastructure construction can be considered to be almost static once they have been implemented meaning that further energy consumption reduction requirements made here are either not possible or can only be implemented over a long term period. Energy consumption in the day to day operation of a rail network is therefore an important consideration for operators, this is a response to rising energy prices and continued constraints on CO 2 emissions. Energy efficient train control is an appropriate and relatively short term strategy that if widely implemented could provide significant energy savings. Train control usually refers the introduction of a type of Driver Advisory System (DAS), where a pre-defined or real time optimal driving strategy is presented to the driver who is encouraged to follow it to reduce energy consumption while keeping within timetable constraints. However, a simplified approach might be to target driver training and evaluation. An example of this is instruction to use established coasting points or to limit aggressive acceleration and deceleration behavior. This paper focuses on affecting energy efficient train control by using driver performance analytics which can be used to improve training and instruction. Previous work (Ellis, et al., 2015) has been made to identify and to characterize train driving approach on a limited selection of rail journeys. This paper identifies the dominant driver styles in operation on the Northern Line of the Merseyrail Network over a period of three months, with a relatively small driver population, an attempt at defining