Gavin Astin 29 September 2011 Suggestions for Efficient Measures Short and Medium term

Gavin Astin 29 September 2011

Suggestions for Efficient Measures

Short and Medium term

Freight train derailment risk model29 September 2011

Agenda

2

1. Background

2. Quantitative Assessment

3. Qualitative Assessment

4. Conclusions


Next Step – Efficient Measures

Is the benefit secured (avoided derailments) worth the investment?

An investment made will return benefits over the lifetime of the measure. Economic indicators used:- Net Present Value – the difference between the present value of cash inflows and the

present value of cash outflows.- Benefit / Cost Ratio – the ratio of benefits to costs (a ratio greater than 1 indicates that the

benefit outweighs the cost).- Internal Rate of Return - can be defined as the break-even interest rate which equates the

Net Present of a projects cash flow in and out.

It is the consideration of these factors that lead to our assessment of those measures that are the most efficient.

3


Agenda

4

1. Background



4. Conclusions


Potential Benefits

5

MeasureAvoided Fatalities

Avoided Track Damage (km)

Avoided Wagon Damage (number)

Avoided Operation

Disruption (hrs)

Avoided Environmental

EventsDerails

prevented

P1-Check Rail 0.16 35 109 751 3 23P2-Track Lubrication 0.09 20 61 422 2 13P10&12-HABD/HWD 0.47 70 270 1889 8 53P11-BAM 0.41 63 240 1673 7 47P13-WLID/WIM 0.59 104 366 2542 10 74P15-Bogie Hunting Detector 0.29 63 199 1377 5 42P16-Wheel Profile 0.14 30 95 657 2 20P18-Track Geometry 0.36 85 280 1941 7 58P19-Clearance Flange Groove 0.04 6 23 164 1 4.5P28-Roller Cages 0.29 44 169 1180 6 33F6-Anti Lock Device 0.17 28 99 693 3 20F7-Sliding Wheel Detector 0.06 10 35 241 1 7

Severe derailments

saved

M1a-Derail Det All Freight 0.96 341 379 2881 17 76M1a-Derail Det All DG 0.85 45 50 380 4 10

M1a-Derail Det RID 0.12 9 10 76 1 2

Annual Benefits


Quantitative Assessment

6

Based on our assessment of available benefit versus cost to secure that benefit at 20 years.

AD = Avoided Derailment; AS = Avoided Severe Derailment (prevents an initially non severe derailment from escalating in consequence).

Rank Measure 10 years 20 years 40 years 10 Years 20 Years 40 Years 10 years 20 years 40 years

1 P13-WLID/WIM 379 756 1,183 3.1 5.1 7.4 51% 52% 52% 74 AD2 P28-Roller Cages 109 284 482 1.7 2.9 4.2 16% 21% 21% 33 AD3 P15 Bogie Hunting Detector 80 283 514 1.4 2.2 3.2 8% 14% 15% 42 AD4 P11-BAM 47 294 572 1.1 1.9 2.8 3% 10% 11% 47 AD5 F7-Sliding Wheel Detector -0 35 75 1.0 1.6 2.4 0% 7% 9% 7 AD6 M1a-Derail Det RID -2 17 39 0.9 1.5 2.2 -2% 6% 8% 2 AS7 P16-Wheel Profile -27 65 170 0.8 1.4 1.9 -4% 5% 7% 20 AD8 M1a-Derail Det All DG -44 56 170 0.8 1.3 1.8 -6% 3% 6% 10 AS9 M1a-Derail Det All Freight -385 303 1,094 0.7 1.2 1.7 -7% 3% 5% 76 AS

10 P10&12-HABD/HWD -507 -257 27 0.5 0.7 1.0 -16% -4% 0% 53 AD11 P19-Clearance Flange Groove -20 -34 -49 0.6 0.6 0.6 N/A N/A N/A 4.5 AD12 P18-Track Geometry -373 -568 -788 0.5 0.6 0.6 N/A N/A N/A 58 AD13 P1-Check Rail -701 -635 -559 0.2 0.3 0.4 -31% -14% -6% 23 AD14 P2-Track Lubrication -276 -459 -667 0.3 0.3 0.3 N/A N/A N/A 13 AD15 F6-Anti Lock Device -3,581 -3,581 -3,580 0.0 0.1 0.1 N/A N/A N/A 20 AD

BenefitNet Present Values Benefit / Cost Ratio Internal Rate of Return


Discussion – Preventative Measures

P10 & P 12 - HABD / HWD:- Extensively applied already, current devices being updated, new devices being installed.- Benefit reducing for this reason.- High installation density required.

P28 - Roller Cages:- Our assessment is for replacement of existing roller cages as soon as possible, an

alternative measure is for replacement at next maintenance interval.

P11 – BAM:- Installation of bearing acoustic monitors along the track that might be an alternative to hot

axle box detectors as less detectors are necessary. Networking of information might be required.

P13 – WLID:- Wheel load and wheel impact load detectors have a broad functionality in terms of detecting

overloading, skew loading, wheel damage and out of roundness which leads to HABD.

A lot of measures directed towards the same cause (hot axle box), but more than 80% of derailments are from causes other than hot axle boxes.


Discussion – Mitigation Measures (with automatic brake application) versus Prevention

Consider measure M1 applied to all Freight trains (M1a-Derail All Freight) and P13- WLID/WIM detectors.

We can see that M1a-Derail Det All Freight prevents 76 derailments from becoming severe whilst P13 prevents 74 derailments from occurring at all.

On first consideration it may seem that preventing 74 derailments is the better outcome. However of these 74 a number will be safely managed (if occurred) and would not escalate in consequence, therefore only a proportion of these prevented derailments are severe (if occurred).

We will see later that M1a has drawbacks which have not been discounted from the benefits. Also the cost of implementing measure M1a means it is less cost effective than P13.

We note an alternative type of derailment detection device which provides an alarm to the train driver when a derailment is suspected, but without an automatic brake application (type M1b). We are not aware of these being available on the market (for freight application) in Europe. Consideration of these devices, the human factors issues involved and their costs would be required prior to formal assessment


Drawbacks

False alarms are possible with most of these measures, although an appropriate monitoring alarm management regime can minimise these.

Item M1 (derailment detection with automatic brake application) could provoke a derailment in false alarm cases.

Measure F7 – Sliding Wheel Detectors are not known to have many suppliers.

9


Agenda

10

1. Background



4. Conclusions


Organisational and Other Options

A lot of our measures could be considered to be replacements or substitutes for proper organisational controls. What other options are there?:

Loading errors can contribute significantly to freight train derailments, usually in combination with other defects like poor train handling / adverse track geometry. - It could be considered to require the qualification and registration of loading personnel.

Track geometry issues, particularly those attributable to track twist, are a major concern. We estimated between 34 and 50 per annum; these include cases where track twist is within existing safety limits, but due to unfortunate freight train composition and loading (which may also be within relevant criteria) combine to cause a derailment: - A system of common and stricter safety limits and intervention limits could be a step forward.

11


Present Obligations

Existing Safety Management Systems should address:- Poor maintenance standards of rolling stock can lead to derailments (including many simple

and routine tasks).- Failure to maintain track geometry and other asset conditions – particularly in secondary and

side-tracks – is a common cause of derailments.

Increased supervision of these parameters by NSAs or other independent parties could be an effective control

12


Agenda

13

1. Background



4. Conclusions


14

Background and Context

We have looked at the potential for improvement, and not an absolute assessment of the efficiency of all measures that are applied today. Therefore it follows that if a measure is applied extensively already there is little room for improvement through the further application of that measure. For this reason some measures that are extensively applied already are not considered in this work. Their omission should not be considered as suggesting such measures are not efficient.

The assessment of measures does not consider the way or the order in which these interventions should be pursued, for example it is not considered whether these interventions should be introduced in a mandatory or voluntary way or whether the measure should be introduced as an EU harmonised measures or only within certain member states or only certain companies.


Efficient Measures (1 to 5)

We consider the following technical measures as being efficient (they have a positive or unity benefit / cost ratio in our reference case and all sensitivity studies):- P13-Wheel Load Impact Detectors / Weighing In Motion.- P28-Replacement of Brass for Polyamide Roller Cages.- P15-Bogie Hunting Detectors.- P11-Bearing Acoustic Monitoring.

Considering measure P28, we have considered an immediate replacement of brass/steel for polyamide roller cages. An alternative option is for replacement at the next scheduled maintenance interval for bearings/axle boxes. This is almost a zero cost option, although the benefits would take longer to materialise.

Potential drawbacks to the use of these measures (excluding measure P28) relate to false alarms. To some extent these can be overcome by the use of good alarm management processes.

15


Efficient Measures (6 to 8)

In addition the following two measures are slightly on the positive side based on the parameters in our reference case but their potential is low and could be eroded by false alarms:- F7-Sliding Wheel Detectors.- P16-Wheel Profile Detectors.

Finally, measure F7 is to the best of our knowledge a market with only a small number of suppliers. This may give rise to market advantage to existing suppliers of these systems.

We consider the following mitigation measure as potentially efficient if the significant identified drawbacks could be solved:- M1a-Derailment Detection (with automatic brake application) applied to All Freight Trains.- (An alternative type of derailment detection device which provides an alarm to the train driver

when a derailment is suspected, but without an automatic brake application (type M1b) may address this drawback. However we not aware of these being available on the market for general freight application. An assessment of these devices, considering the human factors issues involved and their costs, would be required before these could be formally assessed.)

16


Efficient Measures (9 and 10)

F-2: Awareness Programme for Rolling Stock Maintenance. This measure may serve to address the problem of poor maintenance standards of rolling stock. This may include training that sought to concentrate on main rolling stock maintenance derailment causes (which can be extracted from this work). This measure may be followed by increased supervision of these parameters by NSAs.

P-18: Track Geometry (all tracks). The problem of poor track geometry (in particular track twist) should be considered. This is of course an area for each IMs own management system. However a specific measure in this regard must be concerned with increased supervision of these parameters by NSAs to ensure that documented standards are being met.

17


18

End of Session - Any Questions


19

Safeguarding life, property and the environment

www.dnv.com

Documents

Gavin Astin 29 September 2011 Suggestions for Efficient Measures Short and Medium term