Optimum Allocation and Utilisation of track possession time: A case study of tamping. Stephen Famurewa, Tao Xin, Arne Nissen, Uday Kumar

Transcription

1 Optimum Allocation and Utilisation of track possession time: A case study of tamping Stephen Famurewa, Tao Xin, Arne Nissen, Uday Kumar

2 Contents Introduction Research question Railway track behaviour Track geometry quality Tamping strategy Case study Result & Conclusion

3 Introduction to project Train Km (goods) x Train Km (passenger) x Note: Completed Capacity enhancement plans Year 2010 Out of traffic <60% 60-80% % Train Km (goods) Train Km (passenger) Capacity Utilisation (% Line segment) Capacity assessment of Swedish transport administration network Maintenance improvement Framework

4 Introduction to article Railway transport is expected to be competitive and sustainable transportation. Competitiveness requires increase in quantity (capacity) and quality of service. Sustainability ensures tomorrow s competitiveness is not limited by present performance.

5 Research Question Decision support for allocation and utilisation of track possession time (efficiency). Strategy to support quality (effectiveness) and sustain quality of track. Track strategy = Investment strategy + Maintenance strategy Existing infrastructure

6 Railway Track Behavior Patience? During a period of inadequate maintenance, it is a relatively long time before the track shows signs of stress Anger? The track reacts badly & irreversibly to a prolonged maintenance deficiency and could result into undesired consequence Should we understand it then we can manage it- Track geometry parameters map this behavior Conventional or derived parameter

7 Railway track behavior Track geometry quality EN Track quality Estimation/prediction Trend Grey Model Linear Model Exponential Model Randomness Markov Model Fourier Transform Artificial intelligence Smoothening technique

8 Standard dev of long defects Track geometry prediction Good track quality deteriorates slowly while poor one does rapidly b d b dt solving the differential equation e o bt * 2,5 2 Tamping threshold 1,5 1 Deterioration Recovery 0, Days

9 Existing tamping strategies Assumed degradation rates of the track geometry, Known critical/problem spots Track geometry data showing where problems are emerging Availability of the tampers itself Correction of isolated defects (c-failures) Line tamping when TQI becomes very low Advantage of short term saving but long term cost

10 Proposed strategy Line tamping using mulitple sleeper tamper Spot tamping Both are based on longitudinal defect Track possession time Systematic + Spot tamping Fix the root cause of isolated defects and not tamping

11 Case Study Northern section of iron ore line Length considered is ca 134km Divided into 200m segment Renewal period Special treatment for S&C, critical areas (stations) with recurrent c- failures or irregular measmt Track Marker Length (Km) ,4 28,13 0,23 33,34 34, Renewal Year

12 Track geometry measurement data 2,5 Location 2,0 1,5 1,0 Long defect 1529,0 1529,5 1530,0 1530,5 1531,0 1531,5 1532,0 0,5 0, Measurement date inspections from April Aug 2012

13 b with distance 0,0025 0,0020 Deterioration rate 0,0015 0,0010 0,0005 0, Segment from Kiruna to riksgränsen Distribution of b

14 Modelling process INPUT V=tamping speed u=travel speed t p =prepratn time n=nos of segments b=degradatn rate C p C f h p h c p c N np = 1,2,3,...10 N=2years, 730 days MODELLING PROCESS Tamping strategy 1. Direct 2. Distributed N sp, N sc, T.C p, T.T p T.C c, T.T c TC TT OUTPUT

15 Modelling Process Strategy 1: Direct Search for Worse area Each segment is characterised by, b, description of the quality and its evolution over time

16 Modelling Process Strategy 2: Selective Search for Worse segment

17 Optimization Framework Track geometry data Flow chart

18 Model Formulation Constraints & Limits Growth of defects When to tamp Otherwise when Limit for a shift bs ( s, t) e * o t snm 1 p nm nm, t s, t s 1,, N nm ( st, ) c s sd sd s o p o sp d 6 d 200m v u v EN: Recovery or improvement

19 Number of segments Model Formulation, x s 730 N f s t pc f x 0 0 x 0 threshold Objective Functions s 730 Total cos t c * f c * f 1 1 p p c c 1, s, t p 1, s, t c f ps, t, fcs, t 0 else 0 else

20 Optimization i. The problem is a mixed binary non linear programming ii. The optimization is done using FORTRAN. iii. The expected output include

21 Result Estimation of tamping need (nos of segment that will require tamping over a period of time) Optimum number of systematic tamping for a given period of time Number of spot tamping for a given period of time under a specified specified preventive Expected track possession time for different scenario of the parameter combination

22 Conclusion Estimation of maintenance need is facilitated with this approach. The booking of tamping machine is enhanced and planning of tamping action. Desired geometrical quality of track can be supported and sustained in an effective way. Implementing this approach with a proactive way of treating isolated failure will contribute to long term cost saving and reduced track possession time.

23 Future work Introduction of multiple machine parks Emperical study of recovery or improvement. Accomodation of critical spots into the model Consideration of bandwith of exponential growth. Validation of model and improvement.

24

25 THANKS