Download as pdf or txt
Download as pdf or txt
You are on page 1of 30

Three Key Impacts of a Data-Driven Approach

on Your MI Program & Turnaround Strategy

Lewis Makin Ryan Myers


VP, Solutions & Marketing Product Manager
Objective
Discuss how facilities can use their data to:
1. Improve Integrity Operating Window (IOW) Management
2. Optimize Turnaround Scoping
3. Identify the Biggest Threats to Reducing Risks and Costs of
Specific Condition Monitoring Locations (CMLs)

Understand how to leverage quantification of data uncertainty


across your system to improve mechanical integrity and
turnaround performance.

Property of Pinnacle
Perception that inspection dollars ($) are being wasted

• Industry still experiencing significant leaks, particularly on piping.


• How do we identify the right inspections and the right time to inspect?
• How do we justify the required spend and quantify business value?

Hard to justify where inspection is valuable

• How do we decide which inspection locations are valuable?


• How do we confidently defer inspections?
• How do we handle dynamic operating conditions and excursions?

Current industry approaches are highly dependent on expert and user input

• Significant time spent on data “clean-up”.


• Re-inspecting assets with “high confidence” due to uncertainty.
• Unable to quickly process through all of the data to identify real threats.

Property of Pinnacle
The Problem

Consequence
Current POF models:
1. Are deterministic. We do not quantify the
degree of uncertainty within our current
data set.
2. Primarily considers two variables to

Probability
determine a POF: a thickness value and a
degradation rate

Current Risk

Property of Pinnacle
Quantifying Uncertainty

SINGLE CML WALL THICKNESS

Property of Pinnacle
Uncertainty-Based RBI Approach
1. Quantifies uncertainty in data
2. Leverages multiple data inputs to
calculate uncertainty in our model that
could influence our degradation rate
3. Uncertainty modeling allows us to:
• Adjust inspection plans based on actual
operating conditions recommend changes
to improve asset life
• Strategically optimize TAR Strategy –
scope and intervals
• Prioritize which CMLs influence POF
• Identify areas that do not have enough
CMLs
Property of Pinnacle
1 2 3

Improve Integrity Optimize Identifying the


Operating Turnaround Biggest Threats
Window (IOW) Scoping to Reducing
Management Risks and Costs
of Specific CMLs

Property of Pinnacle
Integrity Operating Windows - Overview
Integrity Operating Windows (IOWs) are sets of limits assigned to specific variables that
could affect the integrity and reliability of a process unit.

Critical limit high Failure occurs quickly


IOWs enable owner/operators to more
Failure occurs with sustained
effectively manage risk, optimize cost, Standard level high
operations
and maintain their MI/RBI Programs:
Target Range High
• Improve communication and Stable,
collaboration across departments Reliable, Target – Optimal Safe to operate indefinitely
Profitable
Target Range Low
• Dynamic updates to engineering
models and task plans Failure occurs with sustained
Standard level low
operations
• Rapid response to critical
Critical limit low Failure occurs quickly

Property of Pinnacle
Integrity Operating Windows - Today
• Manually set by Subject Matter Experts
Temperature
(SMEs)
• Monitor critical reliability parameters
• Alarm when a window is breached
IOW Alert • Manual, time-consuming, delayed analysis
if long-term plans should be adjusted
Collect Data Risk Analysis Inspection Planning

Thickness Readings
Current Risk

Property of Pinnacle
Typical Results

High Initial Value and Difficulty in Sustaining Sub-Optimal


Ability to Prevent the Program –> Analysis and
Catastrophic Failures Reduced Effectiveness & Recommendations
Business Value

Property of Pinnacle
Root Issues

Not Integrated with the Not Linked to Business Insufficient Degradation


Risk-Based Inspection Results and Reliability Modeling Capabilities
(RBI) Program Metrics

Property of Pinnacle
Real-Time Degradation Modeling
Temperature

Today

Property of Pinnacle
What-If Analysis
HSE Risk ($, soft,
Today
Degradation Rate ↓ ETC.) ↓

Probability of
Uncontrolled Failure (%) ↓ Lifecycle Cost ($) ↓

Controlled Availability (%) ↑

Property of Pinnacle
Value-Driven Alerting
Temperature
Degradation Rate 50% Increase ↑

Probability of Failure
(%) 10% Increase ↑

Availability (%) 2% Decrease ↓

Lifecycle Cost ($) 30% Increase ↑

Property of Pinnacle
1 2 3

Improve Integrity Optimize Identifying the


Operating Turnaround Biggest Threats
Window (IOW) Scoping to Reducing
Management Risks and Costs
of Specific CMLs

Property of Pinnacle
TAR Planning Today
TAR Planning (Scope & Intervals) differ across owner/operators.
However, there are common themes:
• Preset intervals, budgets, & scope + pressure to push out/reduce
without any negative impacts
• Incorporate outputs from Risk-Based Inspection (RBI) program
• Multiple rounds of qualitative discussions to refine scope
• Scope decisions largely based on current asset health
• Escalate critical items if budget constraint is breached; difficult to
justify the value case and the long-term impact

Property of Pinnacle
Challenges and Root Issues
• Recency Bias: results in conservative inspection and
maintenance

• Qualitative: bartering, trading, and/or whoever is loudest

• People intensive: heavily rely on experience

• External dynamic changes can cause a ripple effect on planning


requiring significant human capital to redevelop the plan

• Ineffective TAR scoping leads to both waste and surprises

Asset health drives significant decisions to determine what is in


and out of scope.
Property of Pinnacle
Apply Uncertainty Models to
Relevant Failure Modes

Property of Pinnacle
Forecast System Availability

Thinning Cracking Fouling

Property of Pinnacle
What If Analysis & Optimization

Property of Pinnacle
1 2 3

Improve Integrity Optimize Identifying the


Operating Turnaround Biggest Threats
Window (IOW) Scoping to Reducing
Management Risks and Costs
of Specific CMLs

Property of Pinnacle
Using Uncertainty Quantification to Drive
Condition Monitoring Optimization

Property of Pinnacle
Statistical Modeling for Assets

Observed Area Unobserved Area


(CMLs)

CMLs
Uninspected How much does my observed data tell me about
areas
the unobserved area?
- If only normal corrosion, quite a bit
- If highly anomalous damage, less so

The total asset performance (POF, lowest projected


thickness, etc.) is ultimately dependent on how the
observed and unobserved areas interact together

Property of Pinnacle
Observed Area:
Which CMLs to Inspect?
Consider CMLs for Inspection if they reduce uncertainty for the system:
- Prioritize CMLs where inspection can affect POF

- Consider change at the system level (roll up CML POFs)

Example:

Changes to CMLs 1/2 produce system change


Changes to CML3 will not change System
=> Prioritize CML1 and CML2

CML3 will matter again in a future inspection

Property of Pinnacle
Unobserved Area: The Influence
of Data on Corrosion Rates
SME assumes:
Gen. Est. Rate
DM Est. Rate
Gen. SME Rate

• 3 mpy normal corrosion rate


DM SME Rate

• 20 mpy anomalous corrosion rate

Data shows:
• Higher/broader normal rate
• Lower anomalous rate

Model adjusts rates to incorporate both


SME and data rate.

We use these rates to estimate damage


on the remainder of asset

Property of Pinnacle
Optimization
We can pull two levers when
optimizing:

1) Which pre-existing CMLs


will we inspect (update
unmitigated POF curve to
mitigated POF curve)

2) Inspect additional area –


tradeoff (potentially
modified) DM corrosion
rate with inspection data Pre-Optimization Post-Optimization
corrosion rate

Property of Pinnacle
Case Study: Value of CML
Optimization on VDU
Objective Optimization Boundary Assumptions Effort
Conditions
• Eliminate CMLs that don’t • Ten Year Evaluation – • Each measurement
which CMLs are • 2 Months
add value costs $350, including
statistically relevant in • 40 hours from Unit
• Add CMLs where damage scaffolding
that time-period Inspector
may be undetected • Measurements not
• Circuit should stay below • 10 Hours from Process
• Deliver Prioritized List of providing value are sunk
risk threshold through Engineer
CMLs across Unit cost
optimization date
• If risk threshold cannot be • CML prioritization is
maintained within 10 based on ability to
years, all CMLs with PoF > reduce uncertainty
0% will be recommended
for inspection.

Property of Pinnacle
Case Study: Value of CML
Optimization on VDU Piping
Value Category VDU

Total Active CMLs 6709

$1.5MM Recommended CMLs to Inspect 1339

saved over 10- Recommended CMLs to Defer 5370


year period
Additional CMLs Recommended 421 (27)

New Total Active CMLs 1760

Property of Pinnacle
Closing Remarks
• Owner/operators are facing increased pressure to improve
Mechanical Integrity & TAR performance.
• This is a challenge due to the nature of siloed data, disparate
programs, and lack of analytics capability to effectively &
efficiently process the data to drive smarter decisions.
• By leveraging the quantification of uncertainty and effectively
integrating advanced analytics into MI programs,
owner/operators can better use their data to:
- Improve Integrity Operating Window (IOW) Management
- Optimize Turnaround Scoping
- Identifying the Biggest Threats to Reducing Risks and Costs of Specific
Condition Monitoring Locations (CMLs)
Property of Pinnacle

You might also like