Contents

Chapter Page
1. ABSTRACT...............................................................................................................................4
2. INTRODUCTION......................................................................................................................4
3. BACKGROUND.........................................................................................................................5
4. ANNULUS PRESSURE MONITORING SYSTEM.........................................................................9
5. PREDICTIVE AI SOLUTION...................................................................................................11
6. IMPLEMENTATION AND INTEGRATION................................................................................14
7. BENEFITS OF PREDICTIVE AI SOLUTION.............................................................................15
8. PRESSURE AUTOMATIC MONITORING AND CONTROL.........................................................15
9. OIL FIELD WELL MANAGEMENT............................................................................................16
10. DATA AGGREGATION AND ANALYSIS................................................................................17
11. CASE STUDIES...................................................................................................................18
12. FUTURE DEVELOPMENTS...................................................................................................19
13. CONCLUSION.....................................................................................................................19
14. REFERENCES......................................................................................................................20

Equations
Equation 1 – Volume Expansion............................................................................5
Equation 2 – Pressure Expansion..........................................................................5
Equation 3 – Pressure Model................................................................................ 6
Equation 4 – Pressure Model Annulus specific.......................................................8

Figures
FIGURE 1 – SEASON GROUND TEMPERATURE........................................................................................................... 7
FIGURE 2 – ANNULUS PRESSURE MONITORING........................................................................................................ 9
FIGURE 3 – ANNULUS PRESSURE MONITORING MODELLING...................................................................................12
FIGURE 4 – ANNULUS PRESSURE MONITORING PREDICTIVE ANALYTICS..................................................................13
FIGURE 5 – ANNULUS PRESSURE MONITORING AND CONTROL...............................................................................16

| 735744461.docx
1 of 22
2 of 22
1. Abstract
The importance of Annulus Pressure Management for safeguarding well integrity becomes even more
pronounced when considering the limitations of cathodic protection and the potential risks associated with
inner casing corrosion. This integrity of the well casing is directly correlating with the live time of a
production well and the benefits company can gain from this well.

Cathodic protection is a pivotal technique employed to shield the outer well casing from corrosion. It relies
on the establishment of a potential difference to ground at the outer casing, which effectively prevents
corrosion by directing any potential corrosion currents away from the casing structure.

However, the inner casing surface, particularly within the annulus region (from A to B to C), presents a
unique challenge. It cannot benefit from cathodic protection because the necessary electrical potential
difference between steel casing to ground can only be established at the outer casing surface, in contact
with the earth. As a result, the inner casing remains susceptible to corrosion.

To counteract this vulnerability, inhibiting noncorrosive fluid is introduced into the annulus is a common
solution to prevent inner corrosion of the casing. This fluid serves as a protective barrier, preventing
corrosion on the inner casing surface by displacing corrosive ground water from ingress.

It's essential to note that this inhibiting fluid may be prone to migration through the casing shoe due to
thermal expansion and the subsequent buildup of pressure damaging the tightness of the casing shoe. If
this pressure is not adequately managed, there's a risk of corrosive salty water from the surrounding ground
infiltrating the annulus. Such an intrusion can lead to the corrosion of the casing wall, jeopardizing the
structural integrity of the well. The loss of well integrity can lead to loss of containment and uncontrolled
release of hydrocarbons into the environment.

The annulus pressure fluctuation is mainly based on thermal expansions and follows thermodynamic rules
driven by temperature differences. Utilizing the existing process data available around a well and process
with AI/ML algorithms can predict the annulus pressure fluctuations in advance and enables the operators to
take countermeasures which ensure that the annulus pressure maintain within a safe range, which
subsequently ensures that integrity of the well.

2. Introduction
Introducing an advanced annulus pressure monitoring and control system, powered by Artificial Intelligence
(AI), represents a remarkable leap forward in the realm of well integrity management. The marriage of
cutting-edge technology with the intricacies of well operations offers a multitude of benefits that extend well
beyond the mere boundaries of routine monitoring. At its core, the AI-driven annulus pressure control
system is a game-changer, capable of not only safeguarding the well's structural integrity but also
substantially prolonging its lifespan. By proactively detecting anomalies and irregularities in annulus
pressure, the system prevents unexpected failures and mitigates risks. This, in turn, leads to reduced
downtime, a decrease in costly maintenance, and an overall boost in operational efficiency. Consequently,
the well's extended longevity translates into increased revenue, as it remains productive for longer periods,
optimizing returns on investment. In this era of evolving technology, the AI-driven annulus pressure
monitoring and control system is poised to redefine the well management landscape, offering a host of

3 of 22
economic and operational advantages that make it a compelling choice for those seeking to enhance well
integrity and revenue streams.

3. Background
The annulus can be considered as confined space, where liquid is contained. This liquid is subject to thermal
expansion which causes pressure rising and dropping. Both extremes, HH and LL lead has impact on the
integrity of the well and shorten the lifetime of the well.

4 of 22
5 of 22
6 of 22
Thermal expansion of liquids can be described using the coefficient of thermal expansion (α), which
quantifies how a substance's volume changes with temperature. The change in volume (ΔV) due to thermal
expansion can be calculated using the following formula:

Equation Error! No text of specified style in document.1 – Volume Expansion

ΔV=V0⋅α⋅ΔT

Where:

 ΔV is the change in volume.

 V0 is the initial volume of the liquid.
 α is the coefficient of thermal expansion.
 ΔT is the change in temperature.

Equation Error! No text of specified style in document.1 – Pressure Expansion

ΔP=V0n⋅R⋅ΔT

Where:

 ΔP is the change in pressure.

 n is the number of moles of the liquid.
 R is the gas constant.
 ΔT is the change in temperature.
 V0 is the initial volume of the liquid.

These formulas help quantify the pressure increase due to thermal expansion during
startup when the temperature in the annulus rises.

The volume V0 is not known but in can be considered as constant as long as the annulus is
tightly sealed by the casing shoe. With this assumption we can reduce the equation to:

Equation Error! No text of specified style in document.1 – Pressure Model

dP = K * dT

7 of 22
K is well specific slope of pressure increase depending on the temperature increase on the annulus fluid.

Environmental impact on annulus fluid.

Figure 1 – Season Ground Temperature

What is visible in from this temperature diagram, is that temperatures below 10 m are independent for
environmental conditions. Below 10 m the earth is following a constant temperature increase of appr. 1
degree per 100m depth down to reservoir temperature.

8 of 22
During operational conditions the hot reservoir fluid is flowing upward and generates a heat transfer from
inner casing through the outer casing to the earth (heat sink).

Under steady stand conditions we can assume a equilibrium how much the fluid is heating up the annulus
pressure.

This phenomenon can be seen on the annulus pressure as well. As long the annulus is sealed, and steady
state well flow, the annulus pressure is almost constant.

Hence the annulus pressure value is available, the pressure increase is predictable by the temperature
increase which will make the annulus pressure.

Equation Error! No text of specified style in document.1 Error! No text of specified style in document. –
Pressure Model Annulus specific

P annulus = k*dT +B

B is well specific and depends on prefilling of the annulus with liquids.

B and K are not defined or can be retrieve with process calculations because there are to many unknowns
and the effort to do these calculations by well is not justified considering the potential inaccuracies.

However, if historical data are available an AI/ML system can learn the K and B factors and build a time and
process-based model for prediction.

To sum all factors up, we should consider following process parameter impacting the annulus pressure as a
basis for building an AI/ML model.

1. Annulus pressure (A,B,C)

2. Ambient temperature
3. Well flow rate
4. Upstream pressure
5. Upstream temperature
6. Choke position
7. Transition time, for transient conditions
8. Optional downstream temperature if available

9 of 22
4. Annulus Pressure Monitoring System

Figure 1 – Annulus Pressure Monitoring

The annulus pressure monitoring system measure the annulus pressures A,B and C.

The numbering of annuli on a wellhead typically follows a standardized and systematic convention. The
annulus closest to the center of the wellbore, often called the "A" annulus, is usually the most critical and
receives special attention. Here's a common numbering system for annuli on a wellhead:

10 of 22
"A" Annulus: The "A" annulus is the space between the production tubing and the production casing. It is
the innermost annulus, closest to the wellbore's center. This annulus is critical for well integrity and is often
the primary focus of monitoring and control.

"B" Annulus: The "B" annulus is typically the space between the production casing and the intermediate
casing or conductor casing, if present. This annulus is further away from the center of the wellbore and may
not always be present, depending on the well's design.

"C" Annulus: The "C" annulus, if applicable, is the space between the intermediate casing and the surface
casing. Like the "B" annulus, the existence of the "C" annulus depends on the well's design and construction.

The numbering system may continue for additional annuli if a well has more casing strings, and each
annulus is typically isolated from the others to allow for independent monitoring and control. The "A"
annulus is of particular importance because it is usually the annulus that can be most readily monitored and
controlled to ensure well integrity and safety.

The annulus pressure is cascaded from A to C, reflecting the depth of the casing shoe. Annulus A have the
lowest depth and therefore need the highest pressure to prevent increase from ground water. They have a
usual pressure range of:

Annulus Pressure range

A 40-80 Bar

B 10-40 bar

C 5-10bar

The pressure range will vary from field to field, and the safe operating range is set by the reservoir team.
Once the pressure is within this range the integrity of the well is ensured. If the pressure moving out of this
range it is an indication that the integrity of the well is compromised, e.g., leakages.

Depending of cost reasons it can be reduced to A and B annulus pressure for low risk wells (e.g. Sweet low
pressure oil wells) For sour gas wells all 3 annulus pressures should be considered to be monitored to ensure
to capture gas outbreaks at earliest stage.

Depending on site conditions these pressure values are connected by the local DCS/RTU. Hence during the
well overwork all installations on the well head has to be removed, wireless solutions are preferred.

11 of 22
The pressure increase does not experience fast variations based on the masses being involved. The
monitoring and transmission of data is therefore in 10 th of seconds or in minutes which ensures a long levity
of the batterie for wireless solutions.

5. Predictive AI Solution
Under item 3 is described which parameters might have an influence on the annulus pressure.

For building the annulus pressure model from up to 7 independent variables are historical data needed to
build the AI model. Preferable multiple well starts are needed to clearly identify the slope rate (K variable)
and pressure constant (B variable) per well. The model is well specific and cannot be transferred to user
wells. Once the model is built from historical data it can be feed with real time data and predict the annuls
pressure based on time based series models, e.g. Arima model, Kalman, etc.

12 of 22
Figure 1 – Annulus Pressure Monitoring Modelling

13 of 22
 Figure 1 – Annulus Pressure Monitoring Predictive Analytics

In the context of wellhead annulus pressure monitoring, the role of Artificial Intelligence (AI) in predictive
analysis is paramount. AI, and more specifically, machine learning algorithms, revolutionize the way we
approach this critical aspect of well management. These algorithms have the capability to learn from
historical data, identifying complex patterns and relationships within annulus pressure parameters. They can
not only

14 of 22
predict potential issues but also offer valuable insights into the overall health of the wellhead. However, the
effectiveness of AI-driven predictions relies heavily on data preprocessing and feature engineering. By
ensuring that the data is cleaned, standardized, and relevant features are selected, the accuracy and
reliability of the predictions are greatly improved. The synergy of AI, machine learning, and meticulous data
handling in annulus pressure monitoring brings about a proactive and data-driven approach, enabling well
operators to anticipate and address issues before they become critical, thereby ensuring the continued
integrity and optimal performance of the wellhead.

6. Implementation and Integration

The implementation of annulus pressure control and its seamless integration into an existing Distributed
Control System (DCS) system offers a range of benefits. To effectively implement this AI solution into well
monitoring systems, careful consideration of strategies is essential. This includes ensuring that the AI
algorithms align with the specific monitoring needs of the wells and can adapt to changing conditions.
Additionally, the integration of AI with the existing monitoring equipment is crucial. This involves connecting
AI modules with sensors, data acquisition systems, and the DCS, enabling the exchange of real-time data
and insights. Real-world examples of successful integration are illuminating. For instance, in the oil and gas
industry, numerous companies have integrated AI-driven annulus pressure control into their DCS systems,
resulting in reduced operational costs, improved safety, and enhanced decision-making capabilities. These
case studies serve as compelling evidence of the tangible advantages of incorporating AI into well
monitoring systems and underscore its potential to revolutionize the management of annulus pressure in
complex well fields.

15 of 22
7. Benefits of Predictive AI Solution
The utilization of predictive AI analysis in annulus pressure monitoring presents a plethora of compelling
benefits. To begin with, its ushers in a remarkable improvement in accuracy. By harnessing sophisticated
machine learning algorithms, AI systems can meticulously examine and interpret vast datasets with a level
of precision that is often unattainable through manual methods. This heightened accuracy translates to a
more dependable annulus pressure monitoring system, enabling early identification of trends, patterns, and
potential anomalies. Consequently, this not only enhances the overall quality of monitoring but also plays a
pivotal role in reducing downtime and minimizing operational disruptions.

In addition to enhanced accuracy and reduced downtime, predictive AI analysis leads to significant cost
savings. By forecasting maintenance requirements and proactively addressing potential issues before they
escalate, AI empowers operators to optimize resource allocation and avert costly well failures. This, in turn,
results in substantial financial benefits for well field management.

Beyond efficiency and cost-effectiveness, predictive AI analysis offers invaluable environmental and safety
advantages. By preventing unexpected well failures and minimizing emissions, it contributes to
environmental sustainability. Moreover, it enhances safety by reducing the risk of accidents associated with
sudden pressure-related issues. In essence, the integration of predictive AI analysis in annulus pressure
monitoring represents a holistic solution that not only improves operational efficiency but also promotes
environmental responsibility and safety within the well field management context.

8. Pressure Automatic monitoring and control

For a fully automated pressure control system a complete skid shall be added to the well head. From a
reservoir (fluid tank) a positive displacement pump pressuring up annulus A until all Annuli have sufficient
pressure. The pressure control valves V25 and V26 are set in the way that they are holding the mid of the
range of the respective Annulus pressure. The pressure in the Annuli should be Pa>Pb>Pc. The release of
the pressure will be to the next annulus which ensures that the pump is utilized at the minimum. In case of
annulus C has high pressure, the release valves will be open and release to the reservoir until the reservoir
is not overfilled. All valves and piping are ¼ inch and would fit in a box which can be remove easily be
disconnection the 3 pipe connections to the annulus. The manual Ansi valves directly at the well head are
not depicted because they are not relevant for the understanding of the functionality but necessary for the
safety of the well head.

16 of 22
 Figure 1 – Annulus Pressure Monitoring And Control

Normal annulus behaviour Abnormal behaviour

17 of 22
The analytics should find out if the annulus have a leak or it will be over pressurized and provide prior to any
damage a prewarning that the operator can intervene manually or the system automatically.

9. Oil Field Well Management

Well field management for annulus pressure monitoring represents a complex yet essential endeavour,
especially when overseeing a fleet of multiple wells. Managing such a fleet requires a holistic understanding
of the interplay between various wells and their annulus pressure dynamics. The challenges inherent in this
task are multifaceted, including the need for real-time data collection, analysis, and timely decision-making.
Automation emerges as a critical solution to address these challenges effectively. Through automated data
collection and aggregation, as well as the use of advanced programming in languages like C and Python, the
management of multiple wells becomes more streamlined and efficient. Nevertheless, one of the most
groundbreaking advancements in well field management is the integration of predictive AI. This technology
is pivotal in forecasting anomalies, trends, and maintenance requirements across the well fleet. It empowers
operators to make informed decisions and proactively address potential issues, ultimately leading to
improved reliability and performance. In the realm of annulus pressure monitoring, predictive AI is a game-
changer, elevating well field management to a new level of efficiency and effectiveness.

10. Data Aggregation and Analysis

In the context of annulus pressure monitoring, data aggregation and analysis play a pivotal role in ensuring
the efficient and reliable operation of multiple wells within a fleet. Collecting and aggregating data from
various wells is the first crucial step in this process. By harnessing the power of programming languages like
Python, you can streamline data collection and integration, facilitating a holistic view of annulus pressure
across the entire fleet. Once this data is consolidated, the next step involves analysing it for trends, patterns,
and anomalies. This involves the application of advanced analytical techniques to identify deviations from
expected pressure profiles, enabling early detection of potential issues. Furthermore, predictive maintenance
is a key application of this data analysis. By leveraging the insights gained from the aggregated data,
predictive maintenance models can be developed to anticipate maintenance needs and prevent costly
failures, thus ensuring the optimal performance and longevity of wells in the fleet. This holistic approach to
data aggregation and analysis is fundamental to enhancing the overall operational efficiency and reliability of
annulus pressure monitoring for multiple wells.
Once the independent process variables are defined which are considered in the Analysis, historical data can
be used to build up the data model. The data can be uploaded into a python program in form of a CSV file.
# Import necessary libraries
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error

# Load the dataset from an Excel file

# Replace 'your_dataset.xlsx' with the actual file name

18 of 22
 data = pd.read_excel('your_dataset.xlsx')

# Assuming the columns in your Excel file are named as follows:

# 'independent_var1', 'independent_var2', 'independent_var3', 'independent_var4',
'independent_var5', 'dependent_var'
# Define independent and dependent variables
X = data[['independent_var1', 'independent_var2', 'independent_var3', 'independent_var4',
'independent_var5']]
y = data['dependent_var']

It is mandatory that the important data are cleaned up and do not contain irrelevant or false data.
Once the data are uploaded a model can be trained. As we have show at item 3 a linear regression model
would be adequate to meet the design requirements of an annulus pressure prediction model.
# Split the dataset into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Create a Linear Regression model

model = LinearRegression()

# Fit the model to the training data

model.fit(X_train, y_train)

Once the best fit model is found and trained it is ready to predict data.
# Make predictions on the test data
y_pred = model.predict(X_test)

# Evaluate the model using Mean Squared Error (MSE)

mse = mean_squared_error(y_test, y_pred)

# Print the model's coefficients and MSE

print("Model Coefficients:", model.coef_)
print("Mean Squared Error:", mse)

# You can use the trained model to make predictions on new data
# For example, to predict the dependent variable for new independent variables:
new_data = pd.DataFrame({
'independent_var1': [value1],

19 of 22
 'independent_var2': [value2],
'independent_var3': [value3],
'independent_var4': [value4],
'independent_var5': [value5]
})
predicted_value = model.predict(new_data)
print("Predicted Value:", predicted_value)

The input for the predictions can be obtained by a time series model analysing the trend of the
independent variables. (ARIMA model)

11. Case Studies

Presenting real-world case studies demonstrating the effectiveness of predictive AI in annulus pressure
monitoring and fleet well management. The OEM will be able to implement the solution cost effectively
without delay of time and further work shops for common understanding. It is in the interest of the EPC or
engineering company to provide such case studies for implementing analytics or digital solutions.

12. Future Developments

Assuming that the annulus pressure is reliable predictable and mainly depending on the temperature change
which occurs mainly during start up. The ramp up rate of a well can be manipulated /slowed down once it
was found that the annulus pressure will go out for the save operation envelope. E.g., if it is known that the
mobilisation time of the operation team needs 6 hours the system can interfere and would slow down the
well start or freeze the choke at this point to give the operators sufficient lead time to reach the site.

13. Conclusion
Annulus pressure management, bolstered by AI solutions, offers a multitude of benefits in the safety of oil
and gas well operations. Firstly, AI-driven monitoring and control systems provide real-time data analysis,
enabling proactive responses to pressure fluctuations and potential issues. This enhances well integrity by
preventing uncontrolled fluid flow and gas migration, minimizing the risk of catastrophic blowouts. Secondly,
AI aids in optimizing well performance by ensuring that annulus pressure is consistently within mechanical
design limits. This extends the operational life of the well, thereby increasing the return on investment.
Moreover, AI solutions enhance safety, as they can quickly detect and respond to anomalies that might
otherwise go unnoticed. In conclusion, the integration of AI into annulus pressure management not only

20 of 22
fortifies well integrity but also enhances operational efficiency and safety, making it a critical and holistic
approach for the modern oil and gas industry. Third the AI solution is more cost efficient and reliable than
any mechanical solution. Despite the accurate prediction of annulus pressure prior to start up, the
intervention is still in hand of the operators, to release the annulus pressure or top up the annulus fluid in
case of low pressure. It should be decided on case-by-case basis if a fully automated system is justified or
not.

14. References
1. Well Integrity
Part 1: ISO/TS16530-1

2. Well Integrity
Part 2: ISO/TS 16530-2

3. A Systematic Approach to Monitoring and Management of Annulus Pressures

SPE/IADC-189412-MS

4. Well Integrity: A Holistic Approach to Annulus Pressure Management for

Subsea Wells
www.gate-energy.com Viking engineering. GAT2004-GKP-2015.09 Sep.2015

5. Annular Casing Pressure Management for Onshore Well

API RP 90-2

6. Casing and Tubing

API SPEC 5CT

7. Hydraulic Fracturing—Well Integrity and Fracture Containment

API RP 100-1

8. Thermodynamics: An Engineering Approach

McGraw-Hill Education. Cengel, Y. A., & Boles, M. A. (2018)

9. Fundamentals of Applied Reservoir Engineering

Gulf Professional Publishing. Chapman, A. J. (2010)

10. www.predictiveanalyticstoday.com/predictive-modeling/

21 of 22
22 of 22