Real Time Optimization of Gasoil Consumption for Railway Transportation

Transcription

1 Real Time Optimization of Gasoil Consumption for Railway Transportation Jorge Melero Corell #PIWorld 2018 OSIsoft, LLC 1

2 About the PROJECT: R&D POC with System PI on LOGITREN for gasoil saving Valencia Polytechnic University for algorithms CETECK for System PI implementation LOGITREN for POC on real stake holder #PIWorld 2018 OSIsoft, LLC 2

3 About LOGITREN Ferroviaria S.A Rail freight company founded in Currently operates 20 trains/week. Main routes: Valencia Madrid, Zaragoza Barcelona. Up to 35% of operational costs are due to gasoil expenditure. #PIWorld 2018 OSIsoft, LLC 3

4 Objectives of the project Develop new HW for capturing position and consumption data of locomotives Develop a consumption management software that can be included in a more general cost management software for the company Create a tool for evaluating the driving efficiency of train drivers, in order to establish a premium policy #PIWorld 2018 OSIsoft, LLC 4

5 Bossiness problems Addressed Dependency on the expertise of train drivers related to the line and diesel consumption. No Real Time view of consumption for each travel. Unknown GPS position of the train during travel from headquarter. #PIWorld 2018 OSIsoft, LLC 5

6 Implementation details Develop Genetic Models to calculate optimal speed graphics based on time and mass of the charge on the train. Develop and install flow consumption meter and gps device on locomotives. Create a gateway program to extract data from the meter and GPS devices and send it to SQL server. (middleware and RDBMS PI Interface) Use PI System to retrieve data, calculate and visualize with PI VISION and advise driver with optimal speed for each track section. #PIWorld 2018 OSIsoft, LLC 6

7 Fuel consumption (l) Model diagrams Pareto curve defines twoobjective (travel time and fuel consumption) optimum solutions. Extreme solutions are discarded for excessive fuel consumption (green circle) or excessive travel time (blue circle). Travel time (min) #PIWorld 2018 OSIsoft, LLC 7

8 Implementation details Architecture of the solution #PIWorld 2018 OSIsoft, LLC 8

9 Implementation details Fuel metering scheme: 1. Fuel filter 2. Air release valve 3. Injection nozzle 4. Overload valve 5. Fuel tank 6. Supply pump 7. Injection pump 8. Fuel meter A 9. Fuel meter B #PIWorld 2018 OSIsoft, LLC 9

10 Implementation details Recording device in the train cabin Locomotive Fuel meters #PIWorld 2018 OSIsoft, LLC 10

11 BEM 500 Display Raw data received from locomotive: : fuel flow, speed, Tª, etc Calculated data (total compsumption, etc..) Train GPS position. Distance and Consumption from last trip. #PIWorld 2018 OSIsoft, LLC 11

12 ON TRAVEL Display Time and weight selectors. Actual speed, consumption and position. Optimal speed with minimal consumption graphic. Objectives: Minimum consumption getting on time to the stops #PIWorld 2018 OSIsoft, LLC 12

13 Event Frames TRAVEL Detailed Event Frame Event Frame Compare table #PIWorld 2018 OSIsoft, LLC 13

14 Final REPORT (Excel Data-Link) Total consumption of travel. Total distance covered. Start and End times. Time stopped and moving. Special events occurred in travel. Real speed versus scheduled from genetic optimal graphics. #PIWorld 2018 OSIsoft, LLC 14

15 WEB REPORT MANAGER benefits for LOGITREN: DATA-Link reports automation and distribution Manage users, groups and distribution lists. [AD compliance] Send monthly production reports, daily performance, etc. You choose the periodicity. Optionally, validate the reports before they are sent. Validation Workflow (Electronic signature / optional) Connect with external BBDD, ERPs, CMMS, etc [AF connection or direct ODBC] #PIWorld 2018 OSIsoft, LLC 15

16 Results and Business Impact Independency on the experience of train drivers related to consumption. Real time view of consumption Known GPS position of the train during travel Policy grants to best drivers REPORTS with the travels 15% of expected fuel saving (based on historic analysis). $$$$$ and galons. (Example: 1route 1.000Km 375 per day/train > (1 year 4 trains) Saving!!!!* 1,7 Millions of tones of CO2 SAVED!! Fuel is the 35% of global costs of a railway company #PIWorld 2018 OSIsoft, LLC 16

17 LOGITREN Real Time Optimization of Gasoil Consumption for Railway Transportation CHALLENGE To have Real Time data from the train to optimize fuel consumption SOLUTION Develop genetic models to calculate optimal speed graphics based on arrival time and mass of the train. IIoT HW and System PI. RESULTS Real time fuel consumption and GPS data. Two software packages: One for drivers, one for managers. To have Real Time data: fuel, Gps,.. Independency of train drivers experience. User-friendly system to assist drivers. Fuel saving New RT IIoT HW to gathering data on train System PI (AF, PI VISION, Data Link) to process and visualize data Genetic algorithms, fuel consumption and train dynamic models. 15% of expected fuel saving Independency on the experience of train drivers Drivers assistance software based on actual train mass and remaining travel time, to optimize fuel consumption. Manager software to control position and fuel consumption of trains in real time. #PIWorld 2018 OSIsoft, LLC 17

18 Genetic algorithms Metaheuristic model inspired by the process of natural selection. Used to solve optimization problems Achieve sub-optimal solutions with incomplete information and/or limited computation capacity. Fitness function Fitness function Population of candidate solutions (randomly generated) More fit individuals are selected Their genome (defining characteristics) are modified Iterative process: Several generations until optimization is achieved #PIWorld 2018 OSIsoft, LLC 18

19 Fuel consumption (l) Model diagrams Pareto curve defines twoobjective (travel time and fuel consumption) optimum solutions. Extreme solutions are discarded for excessive fuel consumption (green circle) or excessive travel time (blue circle). Travel time (min) #PIWorld 2018 OSIsoft, LLC 19

20 FUTURE Challenges: More screen information: PI-Notifications/event frames for limits upsets notification ALARMs PI-VISION Screens ADVANCED Mathematics for: Predictive maintenance TWIN Digital of Train behavior Early Fault Detection by differences between expected and real behavior #PIWorld 2018 OSIsoft, LLC 20

21 Real Time Optimization of Gasoil Consumption for Railway Transportation Jorge Melero Corell Manager of Studies and Projects Logitren Ferroviaria S.A. Pablo Martinez Fernandez Researcher. Phd UPV (Polytechnic University of Valencia) Ernesto Bedrina Ramirez CEO Grupo CETECK #PIWorld 2018 OSIsoft, LLC 21

22 Questions? Please wait for the microphone State your name & company Please rate this session in the mobile app! Search OSIsoft in your app store #PIWorld 2018 OSIsoft, LLC 22

23 #PIWorld 2018 OSIsoft, LLC 23