Direct Use of Base Failure Data

The QRA study on onboard hydrogen storage (e.g. a hydrogen private car) and hydrogen road transports (e.g. LH2 tanker truck) are focused on the loss of containment of the hydrogen due to either an accident breaching a transport container or a failure of associated equipment such as relief valves or fittings while in transit. The CCPS [ ] quoted that accident-initiated releases tend to dominate the risk of hazardous material transportation. The incident release frequencies of hydrogen road transport and onboard hydrogen storage are more likely to occur due to the traffic accidents than through the system malfunctions. Therefore, the losses of containment frequencies for these systems were estimated directly from the accident rate data. Fault tree analyses were not performed for these systems. Similar to this, the loss of containment frequencies of a hydrogen pipeline may also be estimated by direct use of the base failure rate data. In the following section, these methods are described.

4.4.2.1 Estimation of accident frequency for the onboard hydrogen storage

The release chance of an onboard storage is generally a function of the distance travelled [126]. Thus, the frequency of an accident is often expressed as an accident rate per km. Accident rates of a road hydrogen transport for a given road length may be calculated from historical data records according to the equation:

Accidents number per year Accidents . „ _

Vehicle - km per year vehicle - km

A typical value for the road accident rate is 3.0 x 10-6 accidents/vehicle-km [126]. The loss of containment (release) frequency for the onboard hydrogen storage can be estimated from the above accident rates, by using the following equation:

where:

F = Expected frequency of hydrogen release from onboard storage(1/yr)

Fa = Accident rates (1/veh.km)

La = Annual distance for a given vehicle (km/yr)

Pr = conditional probability for release given an accident

4.4.2.2 Estimation of accident frequency for the road transportation

As described in Section 4.3.3.2.1, the release frequency of a truck transport is dominated by truck accident such as collisions and overturning. The non-accident-initiated release or railroad grades do not contribute substantially to the risk analysis. Therefore, the release frequency of the hydrogen road transport can be estimated from the accident rate, as given in the following equation [3]:

where

F = Expected frequency of hydrogen release from road transport(1/yr)

Fa = Accident rates (1/veh.km)

Ls = Length of the route segment (km)

T = Truck per year travelling on the route (1/yr)

Pr = Conditional probability for release given an accidents (-)

It is similar to the previous section, with exception that the annual distance of the truck is calculated for a certain route. The truck accident rates (Fa) can be obtained from the truck road accident data (see Table 5-16). As for comparison, truck accident rates (in accidents per veh-km) for California, Illinois, and Michigan is shown in Table 4-2. The table shows truck accident rates for different broad classes of route types (e.g., urban versus rural, divided versus undivided highway). It also shows that the conditional probability of release of an accident involving a truck carrying hazardous material is a function of vehicle characteristics and the nature of the accident. This conditional release probability has a significant influence on overall risk since it typically addresses the relative likelihoods of different sizes of releases having substantially different potential consequences. With respect to the nature of the incident, the most significant factor is the general accident type.

In the QRA, the transport routes may be divided into several segment routes crossing areas in different population densities (Pop./km2) along the segment routes. For example, in the study two segment routes were considered for the LH2 Tanker truck to deliver hydrogen from a production plant to a hydrogen fuelling station in a city.

Table 4-2 Truck accident rates for California, Illinois, and Michigan [3]

Highway Class

Truck Accident Rate

Conditional

Area

Roadway

(1/106 veh.km)

Probability

Rural

Two-lane

1.36

0.09

Rural

Multilane, undivided

2.79

0.08

Rural

Multilane, divided

1.34

0.08

Rural

Freeway (limited access)

0.40

0.09

Urban

Two-lane

5.38

0.07

Urban

Multilane, undivided

8.65

0.06

Urban

Multilane, divided

7.75

0.06

Urban

Freeway (limited access)

1.35

0.06

Urban

One-way street

6.03

0.06

4.4.2.3 Estimation of failure frequency for a pipeline

Failure frequencies for hydrogen pipelines can be estimated from the existing historical failure rate data of general gas pipeline. For example, the failure rates for the USA compared with those for Europe as shown in Table 4-3. It shows that the pipelines with a diameter of 6 inches or less have an overall failure rate of 1.24 x 10-3/km -year, which is higher than the overall rate for all diameter pipelines. Additional data for gas pipeline is shown in Table 4-4.

Table 4-3 Failure rates of gas pipeline for different causes (1/km-yr) [ 3]

Cause

USA

Europe

All pipelines

lines <=6 in All pipelines

External interference

4.2E-04

8.1E-04

4.2E-04

Corrosion

1.2E-04

9.3E-05

1.1E-04

Material/construction defects

1.6E-04

2.1E-04

1.3E-04

Others

6.2E-05

1.2E-04

5.0E-05

Total

7.7E-04

1.2E-03

7.0E-04

Table 4-4 Failure size in gas pipeline by causes (in %) [3]

Corrosion

Interference

Material failure/

Others construction defects

Rupture

23.4

1.2

10.5

7.9

Hole

50.6

1.3

25.0

2.5

Pinhole

26,0

97.5

64.5

89.6

Total

100

100

100

100

Table 4-5 Failure rates of gas pipeline by cause and size (1/km-yr) for Europe

Size

External Interference

_ Material failure/ Corrosion construction defects

Others Total

Rupture

9.8E-05

1.3E-06

1.3E-05

4.0E-06 1.2E-04

Hole

2.1E-04

1.4E-06

3.1E-05

1.3E-06 2.5E-04

Pinhole

1.1E-04

1.0E-04

8.1E-05

4.5E-05 3.4E-04

Total

4.2E-04

1.1E-04

1.3E-04

5.0E-05 7.0E-04

The failure data from Table 4-3 (for Europe) can be combined with the data on failure size by cause from Table 4-4 to produce the failure rates by cause and size given in Table 4-5. For example, in the Table 4-3 gave a failure rate of 4.19x10-4 /km-year for external interference. Using the failure size distribution for external interference from Table 4-4 gives failure rates of 9.8 x 10-5/km-year for rupture, 2.12 x 10-4/km-year for hole, and 1.09 x 10-4/km-year for pinhole. The similar calculation results are shown in Table 4-5.

The release frequency of hydrogen from a pipeline can be calculated using a simple model given by [183]:

where,

= Expected frequency of hydrogen release from GH2 pipeline(1/yr)

= Failure rate of gas pipeline (1/km.yr)

= Length of the route segments (km)

= Exposure factors for the route segments

Living Off The Grid

Living Off The Grid

Get All The Support And Guidance You Need To Be A Success At Living Off The Grid. This Book Is One Of The Most Valuable Resources In The World When It Comes To When Living Within The Grid Is Not Making Sense Anymore.

Get My Free Ebook


Post a comment