Uncertainty Based Allocation (UBA)

HMA Community of Practice Forum

Nov 2018

Zakiah Zakaria
HCA Lead,
Sarawak Shell Berhad

Copyright of Shell International B.V. 1

Presentation Outlines

➢ Production Allocation
➢ Measurement Uncertainty in a nutshell
➢ UBA Principles per API RP 85
➢ Worked Example & Comparison
➢ Proposal for Relative Uncertainty calculation method
➢ Benefits and Challenges

Copyright of Shell International B.V. 2

Production Allocation
◼ The (mathematical) process of assigning portions of a commingled production stream to the sources, which
contributed to the total flow.
◼ Distributes revenue among different equity holders, across single or multiple countries from potentially varying
product value flowing via common facilities.
◼ What constitute a good Allocation system?
◼ Fully traceable & auditable
◼ Equitable
◼ Accurate
◼ Accepted methods of product Allocation in our industry
◼ Pro-rata
◼ Measurement by Difference
◼ Uncertainty Based Allocation
◼ Yield Factor

Copyright of Shell International B.V. 3

 Systematic
Measurement Uncertainty in a nutshell error
◼ Example of taking a repetitive measurement
of a measurand
◼ Plot the readings
◼ Derive the mean and standard deviation
◼ True value and systematic error
◼ Spread of readings and random error
◼ Confidence Level – 68, 95, 99.7% CL
◼ Systematic error – must be corrected (IR)
◼ Random error – quantifiable statistically
-3𝛿 -2𝛿 -1𝛿 Mean +1𝛿 +2𝛿 +3𝛿
◼ Uncertainty – a parameter associated with the True
result of a measurement that characterizes the value

dispersion of the values that could be ◼ E.g. measurement result with mean of100ml and

reasonably attributed to the measurand Standard deviation of 0.1ml will be written as:
100ml +/- 0.2ml at 95% confidence level
Copyright of Shell International B.V. 4
UBA Principles per API RP 85
◼ A method of hydrocarbon allocation in which the relative uncertainties of the measurements are taken into
consideration, including measurements made by each of the allocation meters, by the reference meters, and by
any other instrumentation, the readings from which affect hydrocarbon flow measurement
◼ System imbalance is distributed to each parties according to the likelihood for that party to cause the imbalance
◼ Allocation factor (∝) is a function of production ratio and measurement system uncertainty
◼ The spread of random error at specified confidence level e.g. 95% CL is proposed to be used for the uncertainty
value though the agreeable actual definition may varies in different allocation network

𝜎𝑖2 𝑄𝑖 2 𝜎𝑧 2
∝𝑖 = 2 𝑛
2
+ 𝑗 2. 2 𝑛
𝜎𝑧 + σ1 𝜎𝑗 σ 𝑄𝑗 𝜎𝑧 + σ1 𝜎𝑗2
𝑖

Allocation factor Production input

Copyright of Shell International B.V. Uncertainty input 5

UBA Worked Example
Field A

Field B 𝜎𝑖 2 𝑄𝑖 2 𝜎𝑧 2
Terminal ∝𝑖 = 2 + .
𝜎𝑧 + σ𝑛𝑗=1 𝜎𝑗 2 σ𝑗 𝑄𝑗 2 𝜎𝑧 2 + σ𝑛𝑗=1 𝜎𝑗 2
Field C 𝑖

Field D

Prod (Q) Rel Unc (U) Abs Unc (σ) Abs Unc (σ2) Alloc Factor UBA
Field
(α)
kbd % kbd kbd2 kbd
Field A 5.00 6.00% 0.3000 0.0900 0.5520 4.17
Field B 5.00 3.50% 0.1750 0.0306 0.2177 4.67
Field C 3.50 4.00% 0.1400 0.0196 0.1421 3.29
Field D 5.50 1.50% 0.0825 0.0068 0.0882 5.37
Total 19.00 0.1470 1.0000 17.5

Terminal 17.50 1.00% 0.1750 0.0306

Imbalance -1.50

Copyright of Shell International B.V. 6

Worked Example & Comparison
Allocated prod in diff methods Difference from initial measurement
Prod Rel Unc Force Bal Force Bal (D
UBA Prorate UBA Prorate
Field (Q) (U) (D abs) abs)

kbd % kbd kbd kbd kbd kbd kbd

Field A 5.00 6.00% 4.17 4.61 4.44 -0.83 -0.39 -0.56
Field B 5.00 3.50% 4.67 4.61 4.44 -0.33 -0.39 -0.56
Field C 3.50 4.00% 3.29 3.22 3.11 -0.21 -0.28 -0.39
Field D 5.50 1.50% 5.37 5.07 5.50 -0.13 -0.43 0.00
Total 19.00 17.5 17.5 17.5 -1.5 -1.5 -1.5

Terminal 17.50 1.00%

Imbalance -1.50

Difference from initial measurement

Field A Field B Field C Field D
0.00

-0.10

-0.20

-0.30

-0.40

-0.50

-0.60

-0.70
UBA kbd Prorate kbd Force Bal (D abs) kbd
-0.80

-0.90
Copyright of Shell International B.V. 7
Proposal for Relative Uncertainty calculation method

◼ Class 1 & 2 : To be taken from Validation Report based on standardized error evaluation method using As
Found/As Left values.
◼ Class 3 & 4:
◼ To jointly identify the various type of meter installations and group into different categories (e.g. DP meter
with P&T, DP meter without P&T etc)
◼ To have a predefined Uncertainty Budget Table (UBT) for each meter category
a) To perform actual calculation once and populate into lookup table for different categories, OR
b) To calculate actual uncertainty using As Found/As Left and wherever necessary; pre-assigned values for parameter(s)
without known data (e.g. density fluctuation, installation effect, etc) at agreed interval. Pre-assigned values are
normally based on worst case scenario / different confidence level and can be updated once more information is
available.
◼ The parties are responsible to make it as academic or as practical as long as it is agreeable

Copyright of Shell International B.V. 8

Potential Benefits

◼ Key enabler for equitable allocation in new tie-ins arrangement where allocation meters within the network are of
different accuracy and performance. Difference in design/configuration and technology, wear & tear, drift etc
◼ Focus and attention get directed to what really matters: metering system performance as directly affecting revenue
distribution
◼ Allows for more flexibility for metering concept select during front end phase, where technical feasibility and
economic analysis of full-fledged metering system is questionable
◼ Monitoring of system imbalance and assigned allocation factor with upper/lower control limits help to trigger
genuine drift in performance somewhere within the network that calls for proper investigation
◼ There will be no negative allocation as oppose to force balance method
◼ In longer term; change in operator’s behaviour as imbalance affect everyone and directly motivates the operators
to strive for excellent metering system performance

Copyright of Shell International B.V. 9

Potential Challenges

◼ Complicated measurement uncertainty calculation and potentially difficult to land on an agreeable standard
calculation method across all parties within the network
◼ Considerably ‘untested’ in complicated allocation network hence implementation will be a learning opportunity
therefore continuous improvement is the way to go

Copyright of Shell International B.V. 10

 Questions and Answers

Copyright of Shell International B.V. 11