P I N N A C L E E X E C U T I V E B R I E F

Quantitative Reliability
Optimization (QRO)
We're introducing the next evolution of reliability -
Quantitative Reliability Optimization. Join us as we
transform how the world approaches reliability.

Pinnacle | Data-Driven Reliability | pinnaclereliability.com | 281.598.1330

 T A B L E O F C O N T E N T S

02 Introduction: The Next Evolution of Data-Driven Reliability

02 The Economics at Stake

03 How We Got Here

03 The Current Challenges to How We Ensure Reliability

05 QRO Simplifies Reliability Modeling and Unlocks the Next

Level of Reliability Performance

08 Join Us on This Journey to Make the World More Reliable

Introduction: The Next Evolution of Data-Driven Reliability

When it comes to the basic needs of modern living— and result in increased performance. To achieve the
commodities—we expect them to be available when optimal balance between maximizing the time a
required and sold at reasonable prices. For this to facility is able to operate and the investment needed
happen, complex industrial facilities must be reliable. to achieve that ideal run time, facilities should use
Over the years, the definition of reliability has varied a new approach to reliability called Quantitative
based on the context and people applying the term. Reliability Optimization (QRO). QRO represents the
When talking at a facility, business, or industry level, next leap forward in reliability modeling and will help
reliable has meant something different than when ensure industrial facilities maximize their investments
talking about a specific asset or piece of equipment while improving their reliability.
within the facility. Pinnacle defines reliability as the
ability to operate when desired. In the following executive brief, Pinnacle will
describe the economics of reliability that are at
As the world’s largest industries place an even stake, how we got to the current state of reliability,
bigger emphasis on lowering costs while sustaining today’s challenges, and how QRO will address those
operations, companies strive to ensure that reliability challenges.
and maintenance investments are in the right places

Pinnacle defines reliability as the ability to operate when desired.

The Economics at Stake

In the fourth quarter of 2020, Pinnacle published the global refiners is easily attainable and would translate
first issue of The Economics of Reliability, a quarterly to a savings of $5 billion across the industry.
report that quantifies the impact of reliability on
various industries. The first report focused on the While this report focused on refining, this analysis can
refining industry and provided an example of the be applied to other complex processing industries
financial impact that optimized reliability programs such as chemicals, mining, and water and wastewater
can have on global industries. treatment. When the estimated savings from the
refining industry are translated to other industries,
In the report, Pinnacle found that complex operating there’s an opportunity for industrial facilities to
companies around the world spend roughly $500 collectively save hundreds of billions of dollars
billion annually on reliability, $50 billion of which is annually through optimized reliability programs. To
spent by refiners. The analysis showed that a 10% realize this savings, we must improve our approach to
improvement in optimized reliability spending for managing reliability.

Pinnacle - Quantitative Reliability Optimization | PG. 02

How We Got Here

Fifty years ago, the industrial age gave way to the Optimization, Failure Modes and Effects Analysis
age of lean manufacturing. Mining operations, power (FMEA), Root Cause Analysis (RCA), Turnaround
generation facilities, chemical plants, refineries and Optimization, Process Hazards Analysis (PHA), and
other large, complex processing facilities began Advanced Condition Monitoring were all models
shifting their focus towards reducing downtime while launched during this era. In refining, for example,
optimizing associated reliability spend. During those the reliability (operating when desired) across the
decades, dozens of programs were created and segment increased by nearly 15% from 1980 to 2010.
standardized to drive towards improved reliability. While each of these models added notable value,
Reliability Centered Maintenance (RCM), Mechanical the impacts of their application have taken years to
Integrity (MI), Risk-Based Inspection (RBI), Spare Parts realize.

In refining, for example, the reliability across the segment

increased by nearly 15% from 1980 to 2010. While each of
these models added notable value, the impacts of their
application have taken years to realize.

The Current Challenges to How We Ensure Reliability

Nearly half a century of analyzing assets and making and advanced data analytics. The goal of these new
improvements in reliability has propelled some reliability initiatives is to provide more quantitative
companies to the top of their industry. However, the insights so that facility leaders and reliability
improvements seen from the models listed above professionals can make objective decisions about
have leveled off over the past decade and companies how to invest and improve facility performance.
are beginning to invest billions of dollars in initiatives To successfully apply data-driven models, we must
that will help them achieve the next level of reliability. recognize that there are gaps in traditional models
Many of these investments are focused on better that are preventing our ability to recognize a return
utilization of standardized large-scale programs on investment.

PG. 03 | Pinnacle - Quantitative Reliability Optimization

 These gaps fall into three categories:

1. Analyses focused on individual assets are failing to optimize the system.

Most traditional models and analyses assess the reliability or risk of individual pieces of equipment,
specific functions of that equipment, or specific damage modes. While this insight is helpful when
evaluating the reliability of a single asset, these analyses often produce limited, subjective information,
making it more difficult for facility leaders to quantify how changes in the performance or data of one
asset affects the reliability of their entire facility. Additionally, facility leaders struggle to compare the
results of their investments because they only have basic rules of thumb to quantify their impact on the
system. As a result, facilities often waste money on initiatives that do not actually improve the reliability
of their system, and in some cases, increase the number of failures.

2. Hyper-conservative risk models increase costs without improving performance.

Traditional models have been highly dependent on empirical formulas, industry standards, and user
inputs. While these have formed a solid foundation for analysis, these approaches are highly simplified.
This oversimplification has introduced substantial uncertainty and therefore inherent risk. To account for
this, models and approaches have been designed to be conservative, which has resulted in projected
failure rates and quantified risk levels that are often an order of magnitude higher than actual. While this
does drive action to reduce risk, it precludes comparison of risk across a system, risks in various assets,
and investment versus return. It also results in overspending to address overestimated issues.

3. Static, siloed models fail to leverage existing data.

To varying degrees, all traditional models are developed using a snapshot of data and records and are
used to build a plan for a specified period of time. The common practice is to implement one program
using one model - for example, risk-based inspection (RBI) for fixed assets - then separately implement
another program with another model - for example, reliability centered maintenance (RCM) for non-
fixed assets. Each of these programs drives specific activities and is not formally linked to another
program or live data inputs. As a result, these programs and models miss the insights associated with
other programs and models, and the ways in which data changes over time is qualitative at best. The
result is both overspending on overlapping programmatic recommendations and increased downtime to
address issues that could have been captured with better quantification of data sensitivity.

Many facilities are hoping that new machine learning their decisions. Unfortunately, while reliability leaders
and artificial intelligence advancements may address are regularly making improvements through their
these gaps, but that is not likely. In the context decisions, the human mind cannot take into account
of reliability, machine learning can be applied to millions of inputs and weigh hundreds of possibilities,
specific assets or problems that can be solved by resulting either in overspending or unrealized
identifying aberrations or repeat patterns. However, reliability. By applying advanced data science
machine learning is limited when it comes to principles to traditional algorithms and strategies,
advanced simulation of complex systems. To identify we can get the best of both: minimizing our spend
patterns and correlations that will unearth insights while maximizing our reliability through quantitative
not previously seen, these programs need large decision making.
sample sizes, repeated incidents, and controlled data
inputs. Complex processing facilities exhibit none Together, we need to address these three gaps to
of these. Therefore, reliability leaders are left with realize both improved facility-wide reliability and
expert opinion and traditional programs to inform optimized spend in a compressed timeframe.

Pinnacle - Quantitative Reliability Optimization | PG. 04

QRO Simplifies Reliability Modeling and Unlocks
the Next Level of Reliability Performance

Quantitative Reliability Optimization (QRO) is a new Risk-Based Inspection (RBI) represented leaps forward
approach to reliability modeling that solves the in the reliability community in past decades, QRO is
current challenges of reliability previously listed and the next leap that combines the best of first principle
enables reliability and operations leaders to improve reliability with the best of data science to deliver
and simplify complex reliability decision making. better outcomes.
Just as Reliability Centered Maintenance (RCM) and

QRO is the next leap that combines the best

of first principle reliability with the best of data
science to deliver better outcomes.

QRO was developed by synthesizing and expanding current models, QRO enables reliability leaders to
upon the best elements of each of the current make data-driven decisions in a way never before
reliability models while introducing new data science possible, providing for improved production levels,
and analytical concepts to close the gaps in current safety, and spending performance with statistically-
programs. By expanding upon the best elements of backed confidence.

PG. 05 | Pinnacle - Quantitative Reliability Optimization

 Reliability Model Comparison

*While RAM uses Monte Carlo simulation, it is driven at the asset level by underlying qualitative Mean Time Between Failure (MTBF) assumptions.

QRO enables reliability leaders to make data-driven

decisions in a way never before possible, providing
for improved production levels, safety, and spending
performance with statistically-backed confidence.

Pinnacle - Quantitative Reliability Optimization | PG. 06

 QRO delivers an optimized reliability plan to facility leaders in three ways:

1. QRO links every failure and data point to the overall system.
To optimize a system, an analysis must cover that system. Rather than isolating models on individual
assets or specific failure modes, QRO statistically relates all of the components of a system into one
analysis. QRO’s analysis can cover thousands of assets for a particular facility or many more if modeling
an entire fleet or supply chain. This statistical relation of all the assets enables us to understand how
critical data and specific failure points relate to the overall production or reliability impact of the system.

2. QRO improves failure models by quantifying uncertainty.

QRO performs a data-driven analysis on each failure point in the system so that each asset is
quantitatively modeled for its unique probability of failure looking forward in time. Instead of only
providing a single probability number, QRO applies an advanced model called the Lifetime Variability
Curve (LVC). The LVC forecasts the distribution of failure probabilities given the level of knowns and
unknowns today. As a result, any failure of the system – whether it’s a stalled rotor on a pump or a
leak in the wall of a pipe – can be accurately modeled given the anticipated point of failure and the
uncertainty associated to that point on the failure curve.

3. QRO provides for a dynamic reliability model for the entire facility.
The QRO model is continually updated by relevant data sources so that key changes in those data
points, whether in process, operations, maintenance, inspection, or economics, update the LVC for each
failure point. As a result, facility leadership can then see how changes to their data affect the reliability
of their system as a whole. In addition, the recommended data gathering tasks from initial studies like
inspections and operator walkdowns provide additional information used to fine-tune the model’s
predictions. These dynamic updates and the ability to see how they impact the reliability of the system
improves the overall confidence of facility leaders that they are making the best reliability decisions for
their facility.

Combining all equipment, failure points, and critical inspection, operations, and reliability departments
data into one analysis engine provides new insights will be able to see exactly how, when, and where
into facility-wide availability. This drives better previously siloed spending resulted in increased
strategic investment decisions and tactical data- performance and increased measurable confidence in
driven reliability planning. Organizationally, plant forecasting.
management and its supporting maintenance,

With QRO, a facility leader or reliability expert can:

• Understand the economic value of every piece of data that is currently being gathered
or possibly gathered in the future, especially from inspection or maintenance activities.
• Drive optimization of all maintenance spending based on strategic reliability targets, in
near real-time, including the implications of moving a turnaround, feedstock changes,
various capital projects, and operating excursions.
• Dynamically plan and schedule tasks, maximizing resource utilization based on value
addition to drive total availability performance.

PG. 07 | Pinnacle - Quantitative Reliability Optimization

Join Us on This Journey to Make the World More Reliable

QRO is the next step in our collective journey to • Ideation

make facilities reliable. QRO synthesizes and expands Thought leadership papers that expand upon the
upon the best elements of current reliability models concepts of what QRO is, how it embodies a novel
while introducing new concepts to drive improved approach, and how it improves upon existing
availability, process safety, and spending performance. reliability models.
As a result, facilities will have greater insight into how
to drive better performance across their system. • Experimentation
Studies to prove the accuracy and credibility of
At Pinnacle, we are explorers. We understand that QRO’s calculations and results, implemented
transformation will not come easily, but we also know across a variety of both theoretical situations and
that transformation is an inevitable part of our journey real-life case studies.
to make the world reliable. We invite you to join us on
our journey through QRO as we establish its capability • Demonstration
to fundamentally transform how we approach We will showcase our toolset for making QRO a
reliability. In the coming months we will be releasing reality.
additional insights into QRO in multiple ways:

Our vision for success is for you to see the power of

QRO and then leverage it within your facility. Join us on
our journey to help make the world more reliable in our
Research and Development Series.

Pinnacle - Quantitative Reliability Optimization | PG. 08