3.06 Reliability Engineering

Every failure is part of a puzzle. The equipment we are maintaining is trying to communicate with use with each and every failure. From alignment errors to lubrication mistakes, to material degradation or wear, there are clues and indications in every failure. And, if we’re paying attention, we can sort out the root cause of the failure along with replacing or repairing the damaged parts. Sometimes though the damage is caused by another issue with the system.

Most maintenance and reliability professionals have seen the six failure patterns (or failure hazard plots), described by Nowlan and Heap. Namely: Bathtub Curve, Wear Out, Fatigue, Initial Break-In, Random, and Infant Mortality. The majority of failures experienced are not directly related to age, but are the result of random or induced failures.   So how does this help when establishing a maintenance program?

R&M professionals are typically the main drivers and beneficiaries of an RCM or similar reliability study at a facility. However, when you invite operations and other key business personnel to participate, we’ve found it often opens their eyes to M&R improvement opportunities and helps paint the picture for future joint improvement efforts. Organizations are then able to operate with the most efficiency, driving toward a world-class reliability program with plant-wide buy-in for the reliability improvement journey. This presentation will discuss a case study of a joint client and partner consultant approach of choosing a machine or line performing below desired performance levels. Using an RCM approach to improve maintenance strategies, the organization experienced reduced downtime and less labour reallocation and idle time, and gained many instant wins like increased visibility in the maintenance budget and increased collaboration between facilities.Presented at MainTrain 2018

The tenets of reliability can be fun and messy. In this workshop, we’ll help you better understand these concepts and where the confusion creeps in. This session will be great for those studying for an exam or for those who want to geek out. You’ll leave with new knowledge, interesting facts, and explainable models that you can take back to your facility. We’ll look at turbo implementations; the criticality of criticality (both of them); the P-F interval; the funky failure curves of RCM; why root cause analysis is a lie; and why predictive tools can’t predict. Bring your own confusions and a willingness to participate in the dialogue as we break each of these down and toss out a few lies.

Asset condition management (ACM) teaches on-condition monitoring for any business with high-capital assets looking to harness machine learning to avoid unexpected failures and control rising equipment maintenance costs. Many businesses are already using continuous condition monitoring technologies like IoT-connected devices. However, beyond simple threshold alerts from condition sensors, extracting real value from the data generated by these sensors for true predictive monitoring requires expert analysis and interpretation. To generate actionable results from condition sensor data, these experts also apply knowledge about the asset’s operation. This limits the value that IoT-enabled ACM can provide to the business. By taking the next step and using advanced algorithms and machine learning to automatically extract real-time insights that drive action, we can now achieve the full potential of ACM. Modern, cognitive online ACM takes data from multiple and varied sources, combines it, and uses AI and machine learning techniques to anticipate equipment failure before it happens. Many reliability professionals recognize the potential of IoT, machine learning, and AI, and are trying to learn these technologies. However, the available training is complex and assumes learners have a background in data science and computer programming. This workshop will provide a beginners’ level understanding of terminology, basic concepts, and techniques to determine how and where you can apply AI in your facilities for meaningful ACM.

In this case study, presented at MainTrain 2017, highlights of an RCM analysis are reviewed including the unexpected outcome. The paper presents a powerful equation derived to calculate the number of inspections required to be performed within the interval between potential and functional failure. Sensitivity analyses are also performed to demonstrate how changes in certain data points affect the results of the analysis. The case study also demonstrates how the recommendation of the analysis was counter-intuitive to conventional thinking given a unique situation and highlights the importance of operational context. Developing an optimal maintenance strategy often requires a systematic approach that includes a Reliability-Centered Maintenance (RCM) analysis. To be successful, these analyses require involvement from many stakeholders and performing a number of pro-active actions to detect or prevent functional failure. Such actions can be unpopular at times and require a solid partnership between the reliability engineering function and Operations and Maintenance.In this case study, highlights of an RCM analysis are reviewed including the unexpected outcome. When there are no safety or environmental consequences, the decision of whether to do an inspection is based on a cost-benefit analysis. This presentation discusses a case study recently performed during a reliability-centered maintenance (RCM) analysis at Cameco’s Port Hope Conversion Facility. The RCM analysis evaluated the cost effectiveness of partially removing a calciner shell to perform a non-destructive examination (NDE) of the bottom of the shell. The RCM uses a specific equation derived to calculate the number of inspections required to be performed within the interval between potential and functional failure. The equation is generic and can be used for any situation.One purpose of this presentation is to demonstrate the identification of the interval between potential and functional failure and how the equation is used so the audience can replicate the analysis in their own situation. Sensitivity analyses are also performed to demonstrate how changes in certain data points affect the results of the analysis. The second purpose of this presentation is to demonstrate how the recommendation of the analysis was counter-intuitive to conventional thinking given a unique situation and highlights the importance of operational context.  

Application of Lean Six Sigma methodology in the optimization of maintenance execution by using data and facts.    As always, equipment maintainability plays an important role in uptime. Besides the reduction of failure rates, the quick recovery from those failures or the successful execution of scheduled activities makes a considerable difference in availability indicators. The application of Lean tools and Six Sigma analysis contributes to the improvement of maintenance execution by applying the 5 steps of Lean Six Sigma methodology (Define, Measure, Analyze, Implement and Control) and using the tools associated with them. This presentation will discuss Lean Six Sigma theory, basic principles of the methodology and case studies showing the use of tools. Case 1 will illustrate the application of Lean Six Sigma in scheduled preventive maintenance for slurry pumps operating in the oil sands industry. Case 2 will examine how the use of Six Sigma analysis reduced the corrosion rate of tubes in a bank of 12 heat exchangers shell and tube type, which heat diluted bitumen upstream of a distillation tower. Both cases emphasize the importance of using data and facts to make decisions, including front end personnel, and the sustainment of implemented solutions. Presented at MainTrain 2017 

Cigar Lake is Cameco’s newest uranium mine located in northern Saskatchewan. During construction it was decided that a lubrication program needed to be implemented to ensure that critical assets were properly maintained. The mine offers challenges in that there is not just one plant or area to setup. There is a fleet of equipment both underground and surface with mobile and stationary assets. In addition there is diesel power generation and a fleet of freeze compressors installed. Each area presents its own challenges and opportunities when setting up a program.There are several aspects of a lubrication program that need to work together to ensure reliability. This presentation will share Cigar Lake’s journey from ground zero towards a world class lubrication program, one that was featured in Machinery Lubrication’s 2016 Lube Room Challenge edition.Why a lubrication program is needed will be discussed. In addition, the improvements made to program management, storage and inventory management, cleanliness, product standardization and sampling will be presented. Lastly, some of the specialized assets in use at the mine will be highlighted and discussed on how they fit into the program. 

A practical and efficient condition based maintenance (CBM) program is built of two fundamental elements-one for rotating and dynamic assets in the adopted term of a (PdM) or Predictive Maintenance program and one for static assets known as Risk Based Inspection (RBI) program. This presentation will highlight these NSPI asset management programs for monitoring the state and reporting condition based (CBM) deficiencies on our physical assets. The discussion will review the development, management, integration and day-to-day application of our PdM/RBI programs including tools and techniques for VA, IR, MCA, HEP/FAC inspections and more. Participants will also discuss the general journey to condition monitoring for equipment over the thermal fleet. NSPI consciously chose innovative vendors, technologies and techniques. The audience will learn what challenges it faced internally and externally. The differences between how the CBM programs are integrated into our business today as opposed to piecemeal prior to AM program implementation starting in 2012.

In an effort to increase equipment reliability and reduce unscheduled downtime, many organizations have taken the proactive step of implementing a Predictive Maintenance (PdM) Program. Unfortunately, only an estimated 20% of these initiatives actually achieve the anticipated results.  This presentation will explore how to avoid the ten most common pitfalls substantially improves PdM results and provide participants with tools they need to implement a best practice preventative maintenance program.