Quantitative risk assessment of crude oil pipelines 1

[Chemister]

Currently, I am working on a project that necessitates a quantitative risk assessment of crude oil pipelines.

I have collected data on crude oil incidents from the CONCAWE cross country European oil pipeline. The database includes incidents for 3 types of service: crude, refined products, and heated black products (hot oil) over 1971-2021. However, I have used the data from 2001-2002 because as this period includes pipelines that were not previously accounted for in their active pipeline network inventory, making it more representative.


The incidents in the database provide information such as pipe diameter, age of pipe, gross volume and net volume, the facility type (undeground, above ground, and pump station), method of detection, the cause of the spill (mechanical, operational, corrosion, third-party activity, or theft), and contamination in the land (m2) and water (yes/no).

Using the pipeline length and assuming that the age distribution for crude oil pipelines follows the same profile as the total pipeline network (crude, refined products, Hot oil), I calculated the weighted mean age for each year and the pipeline exposure, which is the product of pipeline length and exposed duration. I realized that the crude oil pipeline length remained nearly constant (10,000 km) during the considered period, while the pipeline itself progressively aged over time.

Considering the progressive aging of the pipeline, it is possible that aging could be the contributing factor to the absence of a clear trend in the number of incidents and the failure frequency (calculated as the number of incidents divided by the pipeline exposure) with time?. To remove the impact of random fluctuations, I had to employ a moving average technique.
[URL unfurl="true"]https://res.cloudinary.com/engineering-com/image/upload/v1711805924/tips/plHegLrzXl3zWi-Ne0nC_xR73sPNggX_U0vvdHmwqzmgNPnUBI_shufbg.webp[/url][URL unfurl="true"]https://res.cloudinary.com/engineering-com/image/upload/v1711805933/tips/FgjwDxuaFksUkIntiJWS1npwFf5jMAi3Whs7SAD5wGTe_BC_U4_wlmyth.webp[/url]

 

[LittleInch]

There are relatively so few failures that is impossible to get any trends.

You would only expect corrosion to be impacted by age of the pipeline.

Usually you find 3rd party interference is your key risk of significant failure as opposed to pinhole leaks

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Chemister]

@LittleInch Thanks for your response. Actually the data supports exactly what you said that no hole (ancillary equipment failure eg. valve) (8 events), rupture (7 events) and pinhole (7 events) are the significant risks of failure compared to split ( 1 event) and fissure (5 event). Unfortunately, the available data is limited: the CONCAWE report from 2001 to 2021 recorded a total of 48 crude oil pipeline incidents, with hole size data available for only 29 out of these 48 incidents. Utilizing this data, I calculated pipeline exposure for each year from 2001 to 2022 to determine the failure frequency, failure frequency per cause and leak size. Most of the figures generated show no overall trend.

Based on my analysis, I have drawn certain conclusions from the data. For instance, corrosion is the primary cause of pinhole leaks (small-sized leaks), while excavation works are responsible for ruptures. However, I am uncertain about how to further utilize this data beyond reaching these conclusions.

 

[LittleInch]

Look at Bs PD 8010 part 3.

However the key thing that needs specialist action is the consequence or fire and explosion modelling. That needs specialist software.

What are you trying to do exactly?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Chemister]

I intend to apply Quantitative Risk Assessment (QRA) methodology to crude oil pipelines. The methodology involves identifying potential release sources and assessing the frequency and consequences of various incident scenarios. I found a paper with a case study but I have to come up with something new so I was thinking about altering the assumptions used in the paper.

The pipeline segment considered was a buried crude oil pipeline.

I think that notable assumptions that could yield unrepresentative results include
1) Assuming a constant flow rate of release for the whole duration of release to the period where valves are not completely closed.
2) The flow rate for the rupture scenario is equal to the pumping rate.
3) The crater formation was not considered;
4)The jet fire was modeled occurring at ground level

I am having a lot of issues understanding their approach. The reason I am using this paper is that it contains information about the pipeline system, including pipeline design, fluid properties, and operating conditions. However, the pipeline failure incident reports lack data on the size of the failures. As a result, the paper used data from the report "performance of European cross-country oil pipelines." by CONCAWE

The report reported spillage distributions by hole size range, referring to them as no hole, ,split, pinholes, fissures, and ruptures. In the paper, they considered four different failure sizes from each range and assigned them the same frequency as reported in the report. I'm wondering why they didn't consider all pipeline failure sizes, where the equivalent hole size diameter falls between 0 and the pipeline diameter. What if pipeline failures occur at different sizes than the ones considered in the paper? Is it appropriate to use data from database on pipeline failures with a different pipeline system?

 

[LittleInch]

"but I have to come up with something new so I was thinking about altering the assumptions used in the paper."

Why?

QRA is a recognised technique to estimate risk to people from a low probability high onsequence event. The data available isn't great but its all anyone has and is used by all parties. In part because the probability data is rather sketchy, the consequence events are usually quite conservative.

One thing is known though that in a rupture scenario, flow often continues at full flow for several minutes before the operators manage to shut the system down. The are no automatic shutoff systems in pipeline because the false alarm rate is too high.

Look at the code reference I gave you, read it and then apply it.

A QEA is no more than an educated guess as to what the risks are, but regulators generally accept them and normally crude oil pipelines are low risk. The environmental impact is usually far worse.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Chemister]

Thank you for providing this resource. Unfortunately, my professor is requiring us to provide something novel within a short time frame. I have read some parts of the standard, and I am wondering what the event tree would look like for the crude oil in the case study I am dealing with.

Given that the the pipeline diameter and wall thickness are 0.3556 and 0.0071 m, respectively. The pipeline is laid underground and the cover depth ranges from 1.7 to 1.9 m. The operating flow rate under normal conditions is 564.4 m3/h. The crude oil is transported at 3.7 MPa and a temperature of 30 °C.transported liquid is crude oil with a density of 880 kg/m3 (29.3 API) and a dynamic viscosity of 0.02 Pa s. The crude oil, as well as the generated vapors, are flammable. The lower and upper flammability limits are 0.9 and 7.0 (% v/v), respectively. The flashpoint is -6 °C, and the autoignition temperature ranges from 10 to 20 °C. The relative density of the vapor ranges between 3 and 5 times the density of air. Additionally, it is mentioned that the True Boiling Point (TBP) curves of this crude oil indicate that 5% of the crude is evaporated at 73 °C.

The failure can occur as a rupture or leak, and I am uncertain about the state of the fluid as it is discharged from the buried pipeline leak or rupture. Will it have enough momentum to displace the depth of cover, which is 1.7-1.9m, causing the portion of the pipe where the rupture or leak is located to be uncovered and release the fluid outdoors? or will the fluid accumulate inside the trench of the pipe?

It is safe to say that if the fluid (gas or liquid) accumulates in the trench, it will start to diffuse to the cover. However, it wont ignite because oxygen is not present and of the properties of crude oil unless it spreads to the air.