Consultancy. Project Delivery. Innovation. Case Study: Refinery Relief & Flare Study Dynamic Simulation justifies legacy relief header, avoiding £millions in capital expenditure Client: Milford Haven Refinery
Consultancy.
Project Delivery.
Innovation.
Case Study: Refinery Relief & Flare Study
Dynamic Simulation justifies legacy relief header, avoiding £millions
in capital expenditure
Client: Milford Haven Refinery
Who are we?
Process engineering consultancy
Specialising in advanced modelling techniques, to give answers that
are more accurate, reliable and robust
Expertise in HYSYS Dynamics and Dynsim
Decades of process and mechanical engineering experience
Combine cutting edge simulation with real world understanding, to
provide you with solutions that are effective and practical.
Core team of eight engineers, supported by a wide network of
associates.
Extensive simulation experience in a range of industries
Page 2
Project Background
Refinery 130,000 bpd, built 1973, upgraded several times over the years
Following recent significant upgrade needed to demonstrate that original
relief header still adequate
Capital cost of required upgrades could be significant
Disruption of changing existing relief header would affect production
Increasing focus from HSE regulator following some unit upsets
General lack of original data for relief and flare case justifications
Earlier studies were very checklist based – not delivering
Traditional approach to sizing flare headers is likely to be overly
conservative and may not properly understand sequence of events
Page 3
Our Approach
Build high fidelity dynamic simulation models of refinery units
Better understand number of unknowns and assumptions to maximise
existing design contingency
Link all models together to assess interconnected effects
Run agreed set of Scenarios based on a set of agreed, conservative,
credible assumptions
Document work and make efficient recommendations
Page 4
Extent of Study
Accurate, integrated models of key refinery units built from actual refinery
drawings and using plant data.
Crude Tower / Vacuum Tower and Preheat Train
Debutaniser, Virgin and Cat Naphtha Splitters
LPG Train (Depropaniser, Debutaniser)
Naphtha Reformer
Isomerisation
Kerosene Hydrotreater
Diesel Hydrotreater
Ancillary systems (Fuel Gas, Sour Water, Amine)
Page 5
Page 6
Building Models
Data Gathering
Model Building
Model Rating
Testing and
Reporting
We visit our clients, and obtain all the data we need
This includes searching through equipment libraries, reviewing operational trends, and mechanical surveys of pipework and equipment
Detailed input, including all columns, valves, pipe, and pumps
Rigorous models of heat exchangers to accurately simulate heat transfer, pressure drop, and the effects of fouling
We ensure the model matches plant data, from feed through to products.
Site set points and specifications directly compared with models
Once the models are ready, we can test anything the client needs
We supply a comprehensive report of our work, along with a full explanation of all of our recommendations
How we work with our clients
Matching Plant Data – The Traditional Approach
Page 7
Before modelling can commence, the client has to agree a reasonable
set of conditions to match:
Usually done as a point-in-time “snapshot”
Must wait until the whole refinery has been running at steady conditions
for a prolonged period
Can be unrepresentative of normal operation (e.g. operating at unusual
set point)
Susceptible to short term issues, such as ambient conditions, faulty
instruments, or fouling
Matching Plant Data – Our Approach
To avoid the shortcomings of using “snapshots” of data, we
combined this with our own software. This let us:
Page 8
Visualise large sets of
data
Examine months or years
of operation
Amend outliers in
snap-shot data
Identify most common set
points
Understand how units are
run before we began
Model Building
Page 9
Analysis of Results
Page 10
Relief Valve Opening
Frequency on
equipment being
evaluated
Time
Relief Valve Opening
Frequency on
equipment upstream
and downstream of
equipment being
evaluated
Pressure inside
equipment being
evaluated
Example Scenario – Site Wide Partial Power Failure
Page 11
Cumulative Flare Load / Backpressure
Fuel gas PV
Fuel gas Relief
Kerosene Relief
Debutaniser Relief
VNS Relief
Flare
Header
Pressure
A good understanding
of co-incident events
and flow regime in
flare header.
Better understanding
leads to smarter
solutions
Project Outcomes
Page 12
Fully Understand Causes, Timings and Interactions
Site wide recommendations from modelling common mode failures
Preventing simultaneous relief of high pressure and low pressure units
Cost effective solutions to retain current flare network
Recommendations fully checked out
A number of unexpected interactions and events, which were not
intuitive and would otherwise be overlooked
Every piece of equipment (pressure system) evaluated for multiple
relief scenarios in the context of whole operating unit
Avoided need for replacement of legacy flare header
Summary
Page 13
Detailed models built of all refinery units sharing a common flare
header and integrated
All available equipment and process data used to ensure model
accuracy
Hundreds of relief scenarios examined
Fully checked-out recommendations provided to avoid wholesale
replacement of relief header – saving £millions on capital changes
Plant safer and more reliable
Summary
Page 14
Consultancy.
Project Delivery.
Innovation.
Please get in touch to find out how we can help your
business today.
E: [email protected] W: flexprocess.co.uk
T: +44 1454 629 689